1
n Chemometrics‘ and Intelligent Laboratory Systems Chemical Pattern Recognition, by 0. &rouf Research Studies Press, Letchworth, I!18t ISBN 0-86380-044-O This book is published as No. 11 is followed by a section of about 15 in the Chemometrics Series, of which pages on feature selection and a small Dr. Bawden is the Editor. This series section of about 10 pages on visuali- presents reviews and surveys in the zation methods. The next chapter de- field of chemometrics in general, but scribes applications and contains sec- mainly in multivariate data analysis tions on, for instance, analytical data, and structure-activity analysis. The physico-chemical parameters of ele- books are usually state-of-the-art re- ments, chemical structure, spectral views of specific topics, written by data, electrochemical data and chro- specialists, and the general level of matographic data. . the series is good. According to the Introduction, the The present book first describes aim of the book is to summarize re- the methodology of supervised pat- sults achieved from 1979 to 1985. It tern recognition (about 50 pages) and is therefore really an extended and devotes large sections to discriminant critical literature review and not a analysis (including the learning ma- stand-alone book: it has not been chine), K-nearest neighbours rules written for a newcomer in the field and potential methods, modelling methods such as SIMCA and Stroufs who wants to learn about pattern recognition. It contains about 50 own SPHERE method, with smaller pages of references, which means sections on deviation pattern recog- there must be some 400 references. nition and the evaluation of the clas- Full titles are given, which adds to sifiers. the information value. The clustering methodology is de- This review is at the same time scribed in the following 10 pages and very complete and incomplete. It is i, xvi + 202 pages, price f26.65, Chemometrics, by M.A. Sharaf, D.L. Illman and B.R. Kowalski Wiley-Interscience, Somerset, NJ, 1986, XIV + 332 pages, price US$67.20, ISBN O-471 -83106-9 This is basically a textbook that evolved from a senior/graduate course taught at the University of Washington since 1974. It is my be- lief that the authors fulfilled the ob- jective stated in the preface, “M in- troduce the field of chemometrics to advanced students of chemistry”, and highlight the important areas without focusing on mathematical deriva- tions. Each chapter has pertinent refer- ences, suggested readings and exam- ples. There are no problems for the students to work through, however. If there is a shortcoming to this work, I would cite this omission. The chapter on sampling theory is complete in the sense that it does describe nearly all the published work of those scientific workers who move in the field of chemometrics and are more or less known in that circle. It is very incomplete in the sense that it does not cover at all the literature in very related fields such as clinical chemistry, except for a few excur- sions by chemometricians in that area. For the chemometrician, this means that this book is a valuable summary of existing work but that it rarely opens windows on the many other fields, apart from chemistry, in which pattern recognition is applied and where many new solutions to chemometric problems may be found. For the statistician or applied mathe- matician this book may be of’ inter- est, not because of the methodologi- cal developments, but because it gives a good idea of the pattern recogni- tion problems being investigated in chemistry and the methods applied for that purpose. D.L. MASSART VrQe Universiteit van Brussel, Brussels, Belgium excellent. It provides a nice, but brief, introduction to simple statistics and focuses on an important and often overlooked area in analytical chem- istry. Experimental design is a topic that is generally overlooked in the field of chemistry. The topics included in this chapter are all relevant and well covered. The only major omission, Simplex optimization, is covered in a later chapter. The next three chapters, “Signal Detection and Manipulation”, “Cali- bration and Chemical Analyses”, and “ Resolution of Analytical Signals” 253

Chemical pattern recognition, by O. Štrouf : Research Studies Press, Letchworth, 1986, xvi + 202 pages, price £26.65, ISBN 0-86380-044-0

Embed Size (px)

Citation preview

Page 1: Chemical pattern recognition, by O. Štrouf : Research Studies Press, Letchworth, 1986, xvi + 202 pages, price £26.65, ISBN 0-86380-044-0

n Chemometrics‘ and Intelligent Laboratory Systems

Chemical Pattern Recognition, by 0. &rouf

Research Studies Press, Letchworth, I! 18t ISBN 0-86380-044-O

This book is published as No. 11 is followed by a section of about 15 in the Chemometrics Series, of which pages on feature selection and a small Dr. Bawden is the Editor. This series section of about 10 pages on visuali- presents reviews and surveys in the zation methods. The next chapter de- field of chemometrics in general, but scribes applications and contains sec- mainly in multivariate data analysis tions on, for instance, analytical data, and structure-activity analysis. The physico-chemical parameters of ele- books are usually state-of-the-art re- ments, chemical structure, spectral views of specific topics, written by data, electrochemical data and chro- specialists, and the general level of matographic data. . the series is good. According to the Introduction, the

The present book first describes aim of the book is to summarize re- the methodology of supervised pat- sults achieved from 1979 to 1985. It tern recognition (about 50 pages) and is therefore really an extended and devotes large sections to discriminant critical literature review and not a analysis (including the learning ma- stand-alone book: it has not been chine), K-nearest neighbours rules written for a newcomer in the field and potential methods, modelling methods such as SIMCA and Stroufs

who wants to learn about pattern recognition. It contains about 50

own SPHERE method, with smaller pages of references, which means sections on deviation pattern recog- there must be some 400 references. nition and the evaluation of the clas- Full titles are given, which adds to sifiers. the information value.

The clustering methodology is de- This review is at the same time scribed in the following 10 pages and very complete and incomplete. It is

i, xvi + 202 pages, price f26.65,

Chemometrics, by M.A. Sharaf, D.L. Illman and B.R. Kowalski

Wiley-Interscience, Somerset, NJ, 1986, XIV + 332 pages, price US$67.20, ISBN O-471 -83106-9

This is basically a textbook that evolved from a senior/graduate course taught at the University of Washington since 1974. It is my be- lief that the authors fulfilled the ob- jective stated in the preface, “M in- troduce the field of chemometrics to advanced students of chemistry”, and highlight the important areas without

focusing on mathematical deriva- tions.

Each chapter has pertinent refer- ences, suggested readings and exam- ples. There are no problems for the students to work through, however. If there is a shortcoming to this work, I would cite this omission.

The chapter on sampling theory is

complete in the sense that it does describe nearly all the published work of those scientific workers who move in the field of chemometrics and are more or less known in that circle. It is very incomplete in the sense that it does not cover at all the literature in very related fields such as clinical chemistry, except for a few excur- sions by chemometricians in that area. For the chemometrician, this means that this book is a valuable summary of existing work but that it rarely opens windows on the many other fields, apart from chemistry, in which pattern recognition is applied and where many new solutions to chemometric problems may be found. For the statistician or applied mathe- matician this book may be of’ inter- est, not because of the methodologi- cal developments, but because it gives a good idea of the pattern recogni- tion problems being investigated in chemistry and the methods applied for that purpose.

D.L. MASSART VrQe Universiteit van Brussel,

Brussels, Belgium

excellent. It provides a nice, but brief, introduction to simple statistics and focuses on an important and often overlooked area in analytical chem- istry.

Experimental design is a topic that is generally overlooked in the field of chemistry. The topics included in this chapter are all relevant and well covered. The only major omission, Simplex optimization, is covered in a later chapter.

The next three chapters, “Signal Detection and Manipulation”, “Cali- bration and Chemical Analyses”, and “ Resolution of Analytical Signals”

253