Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

Embed Size (px)

Citation preview

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    1/176

    Quantified News based Trading:

    is it the next big thing in algorithmic

    trading ?

    Rajib Ranjan Borah

    Nov 8, 2013PrincetonUChicago Quant Trading Conference

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    2/176

    Agenda

    Background - how is news quantified

    Profitability using quantitative news analysis

    Machine learning techniques for designing quant news strategies

    Q&A

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    3/176

    Agenda

    Background - how is news quantified

    Profitability using quantitative news analysis

    Machine learning techniques for designing quant news strategies

    Q&A

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    4/176

    .

    The world runs on information and few areas as directly so as in

    finance

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    5/176

    Historical Perspective

    1. Rothschild:

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    6/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    7/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    8/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    9/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others largest private fortune in the world

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    10/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others largest private fortune in the world

    2. Reuters:

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    11/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others largest private fortune in the world

    2. Reuters:

    News service used pigeons & telegraph in 1850s to becomefastest news disseminator

    How is news quantified Profitability Machine learning techniques QA

    H i ifi d P fi bili M hi l i h i

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    12/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others largest private fortune in the world

    2. Reuters:

    News service used pigeons & telegraph in 1850s to becomefastest news disseminator

    Continued focus on being the fastest news source

    How is news quantified Profitability Machine learning techniques QA

    H i tifi d P fit bilit M hi l i t h i

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    13/176

    Historical Perspective

    1. Rothschild:

    A family network spread across Europe (Frankfurt, London,

    Paris, Naples, Vienna) enabled obtaining financial

    information before peers

    Knowledge of Battle of Waterloo result one full day before

    others largest private fortune in the world

    2. Reuters:

    News service used pigeons & telegraph in 1850s to becomefastest news disseminator

    Continued focus on being the fastest news source $12.4

    billion conglomerate

    How is news quantified Profitability Machine learning techniques QA

    H i tifi d P fit bilit M hi l i t h i QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    14/176

    Historical Perspective

    How have things progressed since 1850s ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    15/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    16/176

    Historical Perspective

    How have things progressed since 1850s ?

    1850s : Carrier pigeons

    1860s : Telegraph

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    17/176

    Historical Perspective

    How have things progressed since 1850s ?

    1850s : Carrier pigeons

    1860s : Telegraph

    1960s : Teleprinter

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    18/176

    Historical Perspective

    How have things progressed since 1850s ?

    1850s : Carrier pigeons

    1860s : Telegraph

    1960s : Teleprinter 1980s : Electronic network (and internet)

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    19/176

    Historical Perspective

    How have things progressed since 1850s ?

    1850s : Carrier pigeons

    1860s : Telegraph

    1960s : Teleprinter 1980s : Electronic network (and internet)

    2000s : Machine Readable News a.k.a. Quantitative News

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    20/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    21/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    22/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    23/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    How is news quantified Profitability Machine learning techniques QA

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    24/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    q y g q Q

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    25/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    26/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    Apart from trading, quantification of news is also utilized in

    Media evaluation

    Market research

    Brand & reputation management

    Political analysis

    q y g q

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    27/176

    What is Quantitative News Trading?

    Sample output of a News Analytics feed: News

    represented by numbers

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    28/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    Apart from trading, quantification of news is also utilized in

    Media evaluation

    Market research

    Brand & reputation management

    Political analysis

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    29/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    During the 200 milliseconds a human is reading the latest news headline, a

    trading program will have downloaded the entire article, analyzed its

    meaning, & traded based on the content

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    30/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    During the 200 milliseconds a human is reading the latest news headline, a

    trading program will have downloaded the entire article, analyzed its

    meaning, & traded based on the content

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    31/176

    What is Quantitative News Trading?

    News is the first order factor that affects prices, volume,

    volatility of stocks, currencies, commodities, etc

    Computer programs that scan news articles & quantify them

    -> can respond to price moving factors faster than humans

    -> can monitor a vaster amount of news reports than humans

    This field is known as Quantitative News Trading

    During the 200 milliseconds a human is reading the latest news headline, a

    trading program will have downloaded the entire article, analyzed its

    meaning, & traded based on the content

    How do you quantify news reports and articles ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    32/176

    Quantifying News - 1. Sentiment

    News articles are assigned a score called sentiment

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    33/176

    Quantifying News - 1. Sentiment

    News articles are assigned a score called sentiment

    Sentiment says whether the article has a positive / negative or

    neutral tone

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    34/176

    Quantifying News - 1. Sentiment

    News articles are assigned a score called sentiment

    Sentiment says whether the article has a positive / negative or

    neutral tone

    (Sale of Apple iPhones drop = -ve sentiment)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    35/176

    Quantifying News - 1. Sentiment

    News articles are assigned a score called sentiment

    Sentiment says whether the article has a positive / negative or

    neutral tone

    (Sale of Apple iPhones drop = -ve sentiment)

    Sentiment at document level is different from sentiment at

    entity level

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    36/176

    Quantifying News - 1. Sentiment

    News articles are assigned a score called sentiment

    Sentiment says whether the article has a positive / negative or

    neutral tone

    (Sale of Apple iPhones drop = -ve sentiment)

    Sentiment at document level is different from sentiment at

    entity level(Samsung beats Apple in smart phone sales = -ve sentiment for

    entity named Apple, +ve sentiment for Samsung)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    37/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    38/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    39/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    Company Xs sales were good

    Company Xs sales were fantastic

    (Both have one positive keyword, but the intensity differs)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    40/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    Discriminated parser: weighted word count The results were good, not bad.

    The results were bad, not good.

    (Both score equally. Both have the same words - but mean completely

    opposite)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    41/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    Discriminated parser: weighted word count Grammatical parser: which verbs work on which objects.

    check linguistic semantics

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    42/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    Discriminated parser: weighted word count Grammatical parser: which verbs work on which objects.

    check linguistic semantics

    Machine Learning: From the data and the answers, try to find

    the factors

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    43/176

    Quantifying News - 1. Sentiment

    How is sentiment scored ?

    Naive parser: based on word count ofve / +ve keywords

    Discriminated parser: weighted word count Grammatical parser: which verbs work on which objects.

    check linguistic semantics

    Machine Learning: From the data and the answers, try to find

    the factors Generate bag-of-words: distance of subject from these sentiment

    words

    Overfitting (and large vector sets), hitch-hiking and ignorance of

    linguistic structure

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    44/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    45/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing

    A database of words & phrases against which the article is

    searched

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    46/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing

    A database of words & phrases against which the article is

    searched

    Which verbs act on which objects

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    47/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    48/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing issues

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    49/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing issues

    Linguistic structures like negation, double negation, sarcasm,

    intensification, hanging lemma

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    50/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    51/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing issues

    Linguistic structures like negation, double negation, sarcasm,

    intensification, hanging lemma

    (negation: Company X did not become the best in the world

    double negation: Company X did not do bad

    sarcasm: With such an attitude, X is sure to become the best firm

    intensification: Company X did terribly well

    hanging lemma: Company X loses lawsuit against company Y. They will

    have to pay $1billion USD )

    Word Sense Disambiguation - same word, different meanings

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    52/176

    Quantifying News - 1. Sentiment

    Scoring sentiments: grammatical parsing issues

    Linguistic structures like negation, double negation, sarcasm,

    intensification, hanging lemma

    (negation: Company X did not become the best in the world

    double negation: Company X did not do bad

    sarcasm: With such an attitude, X is sure to become the best firm

    intensification: Company X did terribly well

    hanging lemma: Company X loses lawsuit against company Y. They will

    have to pay $1billion USD )

    Word Sense Disambiguation - same word, different meanings Company X received a fine

    X is doing fine X sells fine grained sand, etc

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    53/176

    Quantifying News - 2. Relevance

    Is Sentiment good enough to quantify a news report?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    54/176

    Quantifying News - 2. Relevance

    Is Sentiment good enough to quantify a news report?

    A news article might:

    be predominantly about a company

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    55/176

    Quantifying News - 2. Relevance

    Is Sentiment good enough to quantify a news report?

    A news article might:

    be predominantly about a company

    mention that company and others as well

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    56/176

    Quantifying News - 2. Relevance

    Is Sentiment good enough to quantify a news report?

    A news article might:

    be predominantly about a company

    mention that company and others as well

    mention that company in passing in the article

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    57/176

    Quantifying News - 2. Relevance

    Is Sentiment good enough to quantify a news report?

    A news article might:

    be predominantly about a company

    mention that company and others as well

    mention that company in passing in the article

    Relevance measures how relevant a news article is for aparticular company

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    58/176

    Quantifying News - 2. Relevance

    How is relevance scored ?

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    59/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    60/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    61/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    62/176

    Quantifying News - 2. Relevance

    How is relevance scored ?

    How many companies are mentioned in the news article

    Is the company mentioned in the headline as the

    subject/object

    (Headline:UBS downgrades HSBC is not relevant to UBS)

    In which sentence number is the company first mentioned

    Length of the article & how many times is the firm mentioned

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    63/176

    Quantifying News - 2. Relevance

    How is relevance scored ?

    How many companies are mentioned in the news article

    Is the company mentioned in the headline as the

    subject/object

    (Headline:UBS downgrades HSBC is not relevant to UBS)

    In which sentence number is the company first mentioned

    Length of the article & how many times is the firm mentioned Number of sentiment words & total words in article

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    64/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    65/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    66/176

    Quantifying News - 2. Relevance

    Issues with calculating relevance

    Requires synonym database:

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    67/176

    Quantifying News - 2. Relevance

    Issues with calculating relevance

    Requires synonym database:

    IBM

    International Business Machines

    I.B.M.

    Big Blue

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    68/176

    Quantifying News - 2. Relevance

    Issues with calculating relevance

    Requires synonym database:

    IBM

    International Business Machines

    I.B.M.

    Big Blue

    BAML Bank of America

    Merrill Lynch

    Bank of America Merrill Lynch

    Merrill

    BoA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    69/176

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    70/176

    Quantifying News - 3. Novelty

    Often the news article is not reported in its entirety, but inmultiple spurts

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    71/176

    Quantifying News - 3. Novelty

    Often the news article is not reported in its entirety, but inmultiple spurts

    Alert

    News Article

    Update Append

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    72/176

    Quantifying News - 3. Novelty

    Often the news article is not reported in its entirety, but inmultiple spurts

    Alert

    News Article

    Update

    Append

    Moreover, multiple news

    sources report same news

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    73/176

    Quantifying News - 3. Novelty

    Often the news article is not reported in its entirety, but inmultiple spurts

    Alert

    News Article

    Update

    Append

    Moreover, multiple news

    sources report same news

    News also cause price

    changes which themselves

    become news

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    74/176

    Quantifying News - 3. Novelty

    If we do not keep track & respond to repeated instances ofthe same news

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    75/176

    Quantifying News - 3. Novelty

    If we do not keep track & respond to repeated instances ofthe same news => we will end up repeating our actions

    manifold for the same event

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    76/176

    Quantifying News - 3. Novelty

    If we do not keep track & respond to repeated instances ofthe same news => we will end up repeating our actions

    manifold for the same event

    Therefore every news article should be checked for newnessor novelty before responding

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    77/176

    Quantifying News - 3. Novelty

    How is novelty measured ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    78/176

    Quantifying News - 3. Novelty

    How is novelty measured ?

    The keywords in the current news article are compared to

    historical articles about that company for similarity of digital

    fingerprints

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    79/176

    Quantifying News - 3. Novelty

    How is novelty measured ?

    The keywords in the current news article are compared to

    historical articles about that company for similarity of digital

    fingerprints

    A linked articles count is generated

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    80/176

    Quantifying News - 3. Novelty

    How is novelty measured ?

    The keywords in the current news article are compared to

    historical articles about that company for similarity of digital

    fingerprints

    A linked articles count is generated

    Novelty is reported for

    Within same news feed novelty (i.e. all Bloomberg news articles only)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    81/176

    Quantifying News - 3. Novelty

    How is novelty measured ?

    The keywords in the current news article are compared to

    historical articles about that company for similarity of digital

    fingerprints

    A linked articles count is generated

    Novelty is reported for

    Within same news feed novelty (i.e. all Bloomberg news articles only)

    Across all news feeds novelty (i.e. across Reuters, Dow Jones,

    Bloomberg articles)

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    82/176

    Quantifying News - 4. Market Impact

    Different types of news articles have different impacts on theprice of the asset

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    83/176

    Quantifying News - 4. Market Impact

    Different types of news articles have different impacts on theprice of the asset

    Another aspect of relevance is the likely market impact of the

    news article

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    84/176

    f

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    85/176

    Quantifying News - News Types

    Types of news:

    Accounting news

    Earnings

    Trading updates (broker action, market commentary)

    Guidance

    Financial issues (buybacks, dividends, equity offerings, etc)

    Regulatory filings

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    86/176

    Q if i N N T

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    87/176

    Quantifying News - News Types

    Types of news based on time of news report

    Asynchronous / unexpected

    Synchronous / fixed releases

    Q if i N K F

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    88/176

    Quantifying News - Key Factors

    While the following are the four key inputs:

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    89/176

    Q tif i N K F t

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    90/176

    Quantifying News - Key Factors

    While the following are the four key inputs:

    Sentiment

    Relevance

    Novelty

    Market Impact

    Some news analytics based strategies use other factors as well

    Q tif i N 5 V l

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    91/176

    Quantifying News - 5. Volume

    The number of news articles on the same topic can be a usefulinput to validate the impact

    Q tif i N 5 V l

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    92/176

    Quantifying News - 5. Volume

    The number of news articles on the same topic can be a usefulinput to validate the impact

    Volume of news in Social Media also checked sometimes

    Q tif i N 5 V l

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    93/176

    Quantifying News - 5. Volume

    The number of news articles on the same topic can be a usefulinput to validate the impact

    Volume of news in Social Media also checked sometimes

    News Analytics strategies also check market based qualitative

    parameters along with news -> these help check if reaction to

    news is not already factored in

    Trading Volume in last 24 hours (and historical average volume)

    Price change in last 24 hours

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    94/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    95/176

    Quantifying News 6 Social Media

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    96/176

    Quantifying News - 6. Social Media

    Long term trading strategies try to gauge market sentiment fromthe plethora of information in the social media front

    Search engine volume counts (e.g. Google Trends) - global

    search for news keywords.

    Can be used to confirm market impact of news

    Quantifying News 6 Social Media

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    97/176

    Quantifying News - 6. Social Media

    Long term trading strategies try to gauge market sentiment fromthe plethora of information in the social media front

    Search engine volume counts (e.g. Google Trends) - global

    search for news keywords.Can be used to confirm market impact of news

    Facebook, Twitter - user sentiment evaluated at macro level.

    Quantifying News 6 Social Media

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    98/176

    Quantifying News - 6. Social Media

    Long term trading strategies try to gauge market sentiment fromthe plethora of information in the social media front

    Search engine volume counts (e.g. Google Trends) - global

    search for news keywords.Can be used to confirm market impact of news

    Facebook, Twitter - user sentiment evaluated at macro level.

    Many tools use certified twitter/facebook feeds only

    Quantifying News Key Factors

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    99/176

    Quantifying News - Key Factors

    While the following are the four key inputs:

    Sentiment

    Relevance

    Novelty

    Market Impact

    Some news analytics based strategies use other factors as well Volume

    Social Media

    Quantifying News Market Psyche

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    100/176

    Quantifying NewsMarket Psyche

    News Analytics tools calculate Market Psychology Indices -evaluating broad psychological sentiments from global news

    Quantifying News Market Psyche

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    101/176

    Quantifying NewsMarket Psyche

    News Analytics tools calculate Market Psychology Indices -evaluating broad psychological sentiments from global news

    Country : sentiment, conflict, fear, joy, optimism, trust,

    uncertainty, urgency, violence, government corruption,government instability, social unrest, default, inflation, credit

    tightening, etc

    Quantifying News Market Psyche

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    102/176

    Quantifying NewsMarket Psyche

    News Analytics tools calculate Market Psychology Indices -evaluating broad psychological sentiments from global news

    Country : sentiment, conflict, fear, joy, optimism, trust,

    uncertainty, urgency, violence, government corruption,government instability, social unrest, default, inflation, credit

    tightening, etc

    Equity: Gloom, Anger, Innovation, Stress, Optimism, Earnings

    Expectations, Market Risk, Market Forecast

    Quantifying News Market Psyche

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    103/176

    Quantifying NewsMarket Psyche

    News Analytics tools calculate Market Psychology Indices -evaluating broad psychological sentiments from global news

    Country : sentiment, conflict, fear, joy, optimism, trust,

    uncertainty, urgency, violence, government corruption,government instability, social unrest, default, inflation, credit

    tightening, etc

    Equity: Gloom, Anger, Innovation, Stress, Optimism, Earnings

    Expectations, Market Risk, Market Forecast Currency: Forecast, Currency Peg Instability, Carry Trade

    Quantifying News Market Psyche

    How is news quantified

    Profitability

    Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    104/176

    Quantifying NewsMarket Psyche

    News Analytics tools calculate Market Psychology Indices -evaluating broad psychological sentiments from global news

    Country : sentiment, conflict, fear, joy, optimism, trust,

    uncertainty, urgency, violence, government corruption,government instability, social unrest, default, inflation, credit

    tightening, etc

    Equity: Gloom, Anger, Innovation, Stress, Optimism, Earnings

    Expectations, Market Risk, Market Forecast Currency: Forecast, Currency Peg Instability, Carry Trade

    Agriculture: Acreage cultivated, weather damage, subsidies,

    production volume, supply vs demand, surplus vs shortage,

    price up

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    105/176

    Quantifying News Market Psyche

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    106/176

    Quantifying News Market Psyche

    Agenda

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    107/176

    Agenda

    Background - how is news quantified

    Profitability using quantitative news analysis

    Machine learning techniques for designing quant news strategies

    Q&A

    Is it profitable ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    108/176

    Is it profitable ?

    Are computers smart enough to read news and make profitabletrades?

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    109/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    110/176

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    111/176

    Where Quantified news work

    Machines are faster at responding to events than humans

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    112/176

    Where Quantified news work

    Machines are faster at responding to events than humans

    Machines can process a much vaster amount of information

    without any fatigue

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    113/176

    Where Quantified news work

    Machines are faster at responding to events than humansLow latency event based trading (first to respond)

    Machines can process a much vaster amount of information

    without any fatigue

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    114/176

    Where Quantified news work

    Machines are faster at responding to events than humansLow latency event based trading (first to respond)

    Machines can process a much vaster amount of information

    without any fatigue

    Analyze broad spectrum of news to formulate broad views

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    115/176

    Where Quantified news work

    Machines are faster at responding to events than humansLow latency event based trading (first to respond)

    Machines can process a much vaster amount of information

    without any fatigue

    Analyze broad spectrum of news to formulate broad views

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    116/176

    Where Quantified news work

    Analyze broad spectrum of news to formulate broad views

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    117/176

    Q

    Analyze broad spectrum of news to formulate broad views

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    118/176

    Q

    Analyze broad spectrum of news to formulate broad views

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    119/176

    Q

    Low latency event based trading (first to respond)

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    120/176

    Q

    Low latency event based trading (first to respond)

    For synchronous (fixed releases) expected events (earnings

    releases/ economic figures)

    Company figures provided in xml format instead of text

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    121/176

    Low latency event based trading (first to respond)

    For synchronous (fixed releases) expected events (earnings

    releases/ economic figures)

    Company figures provided in xml format instead of text

    Economic figures provided in binary format instead of textual

    news articles

    Where Quantified news work

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    122/176

    Low latency event based trading (first to respond)

    For synchronous (fixed releases) expected events (earnings

    releases/ economic figures)

    Company figures provided in xml format instead of text

    Economic figures provided in binary format instead of textual

    news articles

    For asynchronous / unexpected news

    Are quantification algorithms robust enough to calculate

    trust-worthy sentiment, relevance, novelty scores ?

    Opportunities : initial under-reaction

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    123/176

    pp

    Quantified news driven trades work even when the trade is doneat the end of the day

    (under-reaction to news immediately. Tetlock, et al)

    Lateendofdayresponsealsoprofitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    124/176

    y p p

    Trading the news immediately = very profitableAt a broad level there is underreaction to news => entering into

    trades at the end of the day also makes profits

    Long short strategy returns

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    125/176

    Filtering sentiments increase profits

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    126/176

    Increasing threshold from 90 to

    95 percentile increases returns

    from 55 to 138 bps in 3 days

    Certain sectors more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    127/176

    Moving from Non-Cyclicals to

    Financials increased the profit

    from 135BP to 147BP

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    128/176

    Small cap firms more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    129/176

    Smaller Cap firms show greater response to extreme sentimentnews event

    (bigger firms have greater scrutiny)

    Filter & trade fewer stocks

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    130/176

    More is not better. Quality over quantity

    Trading only stocks with very high sentiment/relevance is

    better

    Hedged (market-neutral) is better

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    131/176

    Long +ve sentiment stocks onlyOR

    Short -ve sentiment stocks only. Will fail in different regimes

    Being long +ve sentiment stocks & short -ve sentiment stocks

    at the same time gives consistent returns

    Volatility regimes and news

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    132/176

    Volatile vs stable Economic regimes

    In more volatile markets people tend to react less strongly to

    positive news and react more strongly to negative news

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    133/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    134/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    135/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    VIX is low (i.e. surprises during calm times)

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    136/176

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    137/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    VIX is low (i.e. surprises during calm times)

    When markets are improving (i.e. surprise to mostly longposition holders)

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    138/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    VIX is low (i.e. surprises during calm times)

    When markets are improving (i.e. surprise to mostly longposition holders)

    Surprises are more profitable

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    139/176

    Bigger moves happen when there is news in

    Stocks with low beta (i.e. surprises happen to sleepy stocks)

    VIX is low (i.e. surprises during calm times)

    When markets are improving (i.e. surprise to mostly longposition holders)

    Strategy variation - sentiment changes

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    140/176

    Instead of absolute sentiment scores, look at changes insentiment scores of firms

    Bought stocks with highest increase in sentiment

    Shorted stocks with highest decrease in sentiment

    Strategy variation - bottom fishing

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    141/176

    Bottom - fishing / turnaround stories Buying stocks with reversal in sentiment from grossly

    negative (a lot of the stocks turned out to be buybacks)

    Strategy variation - trading volatility

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    142/176

    News articles definitely lead to increased volatility, eventhough direction of move might be difficult to predict

    through news analytics

    Take vega positions (var-swaps) using options in anticipationof increased volatility

    Generating Alpha

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    143/176

    Soft (opinion based) vs. Hard (fact based) newsHard news has a stronger short term reaction than soft news

    Source: RavenPack, FactSet, Macquarie Research, September 2012

    How is news quantified Profitability Machine learning techniques QA

    Generating Alpha

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    144/176

    Scheduled/expected vs. Unscheduled/unexpectedInvestors react more strongly to unscheduled/ unexpected

    news than scheduled/ expected

    Source: RavenPack, FactSet, Macquarie Research, September 2012

    How is news quantified Profitability Machine learning techniques QA

    Generating Alpha

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    145/176

    Forecast vs Actual earningsInvestors react more strongly to forecasts than actual earnings

    news

    Source: RavenPack, FactSet, Macquarie Research, September 2012

    How is news quantified Profitability Machine learning techniques QA

    Generating Alpha

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    146/176

    Guidance vs Actual earningsInvestors react more strongly to guidance to actual earnings

    Source: RavenPack, FactSet, Macquarie Research, September 2012

    How is news quantified Profitability Machine learning techniques QA

    Generating Alpha

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    147/176

    News type Event Study Results

    Source: RavenPack, FactSet, Macquarie Research, September 2012

    To summarize

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    148/176

    News Analytics works best with

    Small cap stocks

    Sectors like pharma, banking, etc

    Stocks with low beta

    When VIX is low

    When markets are improving

    Hard news (vis-a-vis Soft news)

    Unscheduled news events (vis-a-vis scheduled news events)

    Being market-neutral

    Doing fewer stocks, but those with stronger signals

    Quantifying News - Where it fails ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    149/176

    On Sep. 7, 2008 Googles newsbots picked up an old 2002story about United Airlines possibly filing for bankruptcy

    Quantifying News - Where it fails ?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    150/176

    On Sep. 7, 2008 Googles newsbots picked up an old 2002story about United Airlines possibly filing for bankruptcy

    UAL stock dived immediately

    Quantifying News - Where it fails?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    151/176

    News analytics were taught that Osama-Bin-Laden, andkilled had -ve sentiments for the markets

    Quantifying News - Where it fails?

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    152/176

    News analytics were taught that Osama-Bin-Laden, andkilled had -ve sentiments for the markets

    On May 2 2012 when news reporting Osama Bin-Landen

    killed were published, news bots treated this as a negativenews article and sold stocks

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    153/176

    Quantifying Newschallenges

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    154/176

    Languages like Chinese and Japanese with large number ofalphabetic symbols and complex grammar

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    155/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    156/176

    Agenda

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    157/176

    Background - how is news quantified

    Profitability using quantitative news analysis

    Machine learning techniques for designing quant news strategies

    Q&A

    Machine Learning methodologies

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    158/176

    Traditional approach => formulate hypothesis based onexperience/expertise, validate statistically using historical data

    Machine Learning methodologies

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    159/176

    Traditional approach => formulate hypothesis based onexperience/expertise, validate statistically using historical data

    Machine learning approach =>

    Machine Learning methodologies

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    160/176

    Traditional approach => formulate hypothesis based onexperience/expertise, validate statistically using historical data

    Machine learning approach => output + raw data fed into a

    system. System reports factors within data that lead to output

    Machine Learning methodologies

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    161/176

    Traditional approach => formulate hypothesis based onexperience/expertise, validate statistically using historical data

    Machine learning approach => output + raw data fed into a

    system. System reports factors within data that lead to output

    Three broad approaches

    Tree

    Forest

    Planet

    Machine Learning - TREE method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    162/176

    Output: Post-event abnormal resultsInput: Quantitative news analytics

    Machine Learning - TREE method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    163/176

    Output: Post-event abnormal resultsInput: Quantitative news analytics

    Machine Learning - TREE method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    164/176

    Output: Post-event abnormal resultsInput: Quantitative news analytics

    Issues: Overfitting

    (works with training data

    does not work on real data)

    Machine Learning - TREE method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    165/176

    Output: Post-event abnormal resultsInput: Quantitative news analytics

    Issues: Overfitting

    (works with training data

    does not work on real data)

    Solution: Pruning

    Machine Learning - TREE method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    166/176

    Output: Post-event abnormal resultsInput: Quantitative news analytics

    Issues: Overfitting

    (works with training data

    does not work on real data)

    Solution: Pruning

    Machine Learning - FOREST method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    167/176

    Multiple factors might impact output

    Machine Learning - FOREST method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    168/176

    Multiple factors might impact output

    Instead of one tree to solve everything,

    have a forest of trees

    Machine Learning - FOREST method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    169/176

    Multiple factors might impact output

    Instead of one tree to solve everything,

    have a forest of trees

    Machine Learning - FOREST method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    170/176

    Multiple factors might impact output

    Instead of one tree to solve everything,

    have a forest of trees

    Each tree has a vote in the output.

    Weightage of vote depends on accuracy

    of that tree

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    171/176

    Machine Learning - PLANET method

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    172/176

    Instead of linear relationships between input and output,

    Planet breaks the variable space into sections, fits linear

    functions within those sections

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    173/176

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    174/176

    Text Mining: An example

    How is news quantified Profitability Machine learning techniques QA

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    175/176

    Converting a lineof news into

    metadata to be

    used for analysis

    or trade

  • 8/10/2019 Quantified News Based Trading - Princeton UChicago - Rajib Ranjan Borah

    176/176