Abstract. The key chemometric methods and models used to solve The key chemometric methods and models used to solve the problems of qualitative and quantitative analysis and for the problems of qualitative and quantitative analysis and for process analytical technology are considered. The achievements in process analytical technology are considered. The achievements in the field of chemometrics made in the last 20 years are surveyed. the field of chemometrics made in the last 20 years are surveyed. The trends and prospects for its development are discussed. The The trends and prospects for its development are discussed. The bibliography includes 228 references bibliography includes 228 references. I. Introduction 1. The history of chemometrics and its position in the system of knowledge Twenty years have passed since the publication of the Russian translation of the only (until recently) book on chemometrics; 1 much has changed during this period. Currently, chemometric methods are used in various fields of science and engineering. This review is mainly devoted to analytical chemistry where three fields of application of chemometrics can be distinguished: qualitative and quantitative analysis, process analytical chemistry and design of experiments. 2 The attention is focused on the first application, the second one receives less attention and the third is barely considered. This choice of key points is due to the fact that the awareness of Russian chemists of the chemometric methods increases in exactly this order. Numerous papers dealing with design of experiments 3 and metrology 4 are published in Russian scientific journals. The number of publications dealing with chemometrics rapidly grows: 15 years ago *100 papers per year were published, while now their number is more than 5000 a year. Therefore, when preparing the review we reasonably restricted its scope. Analysis of chemical data is the most important direction in chemometrics. In recent years, it has been rapidly and fruitfully developing; analytical chemists have proposed not only new methods for data processing, but also new approaches to experiment setting. Chemometrics is a synthetic discipline at the boundary of chemistry { and mathematics. As is often the case with boundary disciplines, it still lacks a generally recognised definition. The most popular definition was proposed by Massart, 6 namely, `Chemo- metrics is the chemical discipline that uses mathematical, statis- tical and other methods employing formal logic to design or select optimal measurement procedures and experiments, and to pro- vide maximum relevant chemical information by analyzing chem- ical data'. Probably, many people would accept this definition. However, the scope of science should be determined by the objects and goals it pursues rather than by the methods and instruments used. Certainly, the problem of information retrieval from source data is very important for both practice and the development of the theory; however, the experiment setup that would give results containing the required information is equally important. These two equivalent aspects, i.e., retrieval of information from the data and collection of data that contain the desired information, have been reflected in the modern definition of chemometrics proposed by Wold. 7 Chemometrics solve the following problems in chem- istry: Ð how to get chemically relevant information out of meas- ured chemical data; Ð how to represent and display this information; Ð how to get such information into data. The vigorous development of chemometrics in the late 1970s is correlated with the advent, in the same period, of high-speed computer facilities, which have become universally available to scientists and engineers. This allowed implementation of many complicated algorithms, especially for analysis of data obtained in multiresponse and multivariate experiments. As a consequence, more complex equipment capable of performing a much higher number of measurements appeared. However, it turned out that a large amount of data does not necessarily mean that there is enough information. Therefore, analytical chemists have started to use chemometric methods to retrieve this information and to confirm that the conclusions drawn are reliable. This led to the first obvious success. It was found that traditional labour-con- suming analytical methods that require unique equipment and expensive chemicals can be replaced by much faster and less expensive indirect methods. This trend is manifested most clearly in the use of IR spectroscopy, especially in the near region, which has previously considered to be of low utility due to high noise level difficult to eliminate, caused by intense absorption of water O Ye Rodionova, A L Pomerantsev N N Semenov Institute of Chemical Physics, Russian Academy of Sciences, ul. Kosygina 4, 119991 Moscow, Russian Federation. Fax (7-495) 939 74 83, tel. (7-495) 939 74 83, e-mail: oksana@chph.ras.ru (O Ye Rodionova), forecast@chph.ras.ru (A L Pomerantsev) Received 23 August 2005 Uspekhi Khimii 75 (4) 302 ± 321 (2006); translated by Z P Bobkova DOI 10.1070/RC2006v075n04ABEH003599 Chemometrics: achievements and prospects O Ye Rodionova, A L Pomerantsev Contents I. Introduction 271 II. Data and models used in chemical analysis 274 III. Qualitative analysis methods. Exploration, classification and discrimination 276 IV. Quantitative analysis methods. Calibration 280 V. Data preprocessing and signal processing 282 VI. Conclusion 283 { Chemometrics appeared as a separate subdiscipline within analytical chemistry in 1974. 5 B Kowalski (USA) and S Wold (Sweden) can be considered its founders. Russian Chemical Reviews 75 (4) 271 ± 287 (2006) # 2006 Russian Academy of Sciences and Turpion Ltd This article is protected by the copyright law. You may copy and distribute this article for your personal use only. Other uses are only allowed with written permission by the copyright holder.