Alejandro C. Olivieri
Introduction to Multivariate Calibration
A Practical Approach
Gebonden Engels 2018 9783319970967Levertijd ongeveer 9 werkdagen
Gratis verzonden
Samenvatting
This book offers an introductory-level guide to the complex field of multivariate analytical calibration, with particular emphasis on real applications such as near infrared spectroscopy. It presents intuitive descriptions of mathematical and statistical concepts, illustrated with a wealth of figures and diagrams, and consistently highlights physicochemical interpretation rather than mathematical issues. In addition, it describes an easy-to-use and freely available graphical interface, together with a variety of appropriate examples and exercises. Lastly, it discusses recent advances in the field (figures of merit, detection limit, non-linear calibration, method comparison), together with modern literature references.
Specificaties
ISBN13:9783319970967
Taal:Engels
Bindwijze:gebonden
Uitgever:Springer International Publishing
Lezersrecensies
Wees de eerste die een lezersrecensie schrijft!
Inhoudsopgave
1. Chemometrics and multivariate calibration<p>1.1. Chemometrics: what's in a name?</p>
<p>1.2. Univariate and multivariate calibration</p>
<p>1.3. The order and the ways</p>
<p>1.4. Why multivariate calibration?</p>
<p>1.5. Near infrared spectroscopy: the analytical dream</p>
<p>1.6. Multi-way calibration and new advantages</p>
<p>1.7. References</p>
<p> </p>
<p>2. First-order multivariate models: CLS</p>
<p>2.1. Direct and inverse models </p>
<p>2.2. Classical least-squares</p>
<p>2.3. The CLS calibration phase</p>
<p>2.4. Why least-squares? Mathematical requirements</p>
<p>2.5. The CLS prediction phase</p>
<p>2.6. The CLS vector of regression coefficients</p>
<p>2.7. A CLS algorithm</p>
2.8. Validation of the CLS model<p></p>
<p>2.9. Spectral residuals and sample diagnostic</p>
<p>2.10. The first-order advantage</p>
<p>2.11. A real case </p>
<p>2.12. Advantages and limitations of CLS</p>
<p>2.13. Exercises</p>
<p>2.13. References</p>
<p> </p>
<p>3. First-order multivariate models: ILS</p>
<p>3.1. Why calibrating the other way around? A fantastic idea</p>
<p>3.2. The ILS calibration phase</p>
<p>3.3. Mathematical requirements</p>
<p>3.4. The ILS prediction phase</p>
<p>3.5. An ILS algorithm</p>
<p>3.6. The validation of the ILS model</p>
<p>3.7. Advantages and limitations of ILS</p>
<p>3.8. The successive projections algorithm</p>
3.9. A real case<p></p>
<p>3.10. How to improve ILS</p>
<p>3.11. Exercises</p>
<p>3.12. References</p>
<p> </p>
<p>4. Principal component analysis: PCA</p>
4.1. Why compressing the data?<p></p>
<p>4.2. Real and latent variables</p>
<p>4.3. Principal components</p>
<p>4.4. Significant loadings and scores </p>
<p>4.5. Non-significant loadings and scores </p>
<p>4.6. Sample classification with PCA</p>
<p>4.7. Multivariate calibration with PCA</p>
<p>4.8. Exercises</p>
<p>4.9. References</p>
<p> </p>
<p>5. First-order multivariate models: PCR</p>
<p>5.1. Combination of PCA and ILS: another fantastic idea</p>
<p>5.2. Matrix compression and decompression</p>
<p>5.3. The PCR calibration phase</p>
<p>5.4. Mathematical requirements</p>
<p>5.5. The PCR prediction phase</p>
<p>5.6. The PCR vector of regression coefficients</p>
5.7. A PCR algorithm<p></p>
<p>5.8. What is the value of A?</p>
<p>5.9. Advantages and limitations of PCR</p>
<p>5.10. A real case</p>
<p>5.11. What can be better than PCR?</p>
<p>5.12. Exercises</p>
<p>5.13. References</p>
<p> </p>
<p>6. The optimum number of latent variables</p>
<p>6.1. The importance of estimating the optimum A</p>
<p>6.2. Explained variance</p>
<p>6.3. Visual inspection of loadings</p>
<p>6.4. Leave-one-out cross validation</p>
<p>6.5. Cross-validation statistics</p>
<p>6.6. Monte Carlo cross-validation</p>
<p>6.7. Other methods</p>
<p>6.8. The principle of parsimony</p>
6.9. Beyond statistics: physicochemical interpretation of A<p></p>
<p>6.10. Exercises</p>
<p>6.11. References</p>
<p> </p>
<p>7. First-order multivariate models: PLS</p>
<p>7.1. The PLS philosophy</p>
<p>7.2. The PLS calibration phase</p>
<p>7.3. Mathematical requirements</p>
<p>7.4. The number of latent variables</p>
<p>7.5. The PLS prediction phase</p>
<p>7.6. The vector of PLS regression coefficients</p>
7.7. A PLS algorithm<p></p>
<p>7.8. Advantages and limitations of PLS</p>
<p>7.9. A real case</p>
<p>7.10. PLS-1 and PLS-2 models</p>
<p>7.11. Discriminant PLS </p>
<p>7.12. Beyond PLS</p>
7.12. Exercises<p></p>
<p>7.13. References</p>
<p> </p>
<p>8. Comparison of models</p>
<p>8.1. Which is the best model?</p>
<p>8.2. The randomization test</p>
<p>8.3. How the test works</p>
<p>8.4. Algorithm for the randomization test</p>
<p>8.5. PCR, PLS-1 and PLS-2: when, how and why</p>
<p>8.6. Linear and non-linear models: when, how and why</p>
<p>8.7. Tests of multivariate non-linearity</p>
8.8. A real case<p></p>
<p>8.9. Conclusions</p>
<p>8.10. References</p>
<p> </p>
<p>9. Data pre-processing</p>
<p>9.1. Selection of calibration samples</p>
9.2. Calibration outliers <p></p>
<p>9.3. Selection of wavelengths to build the model</p>
<p>9.4. The vector of regression coefficients as selector</p>
<p>9.5. Interval-PLS </p>
<p>9.6. A real case</p>
<p>9.7. Other selection methods</p>
<p>9.8. Mathematical transformation of spectra</p>
<p>9.9. Mean centering</p>
<p>9.10. Smoothing and derivatives: advantages and hazards</p>
<p>9.11. Multiplicative correction</p>
<p>9.12. Other pre-processing methods</p>
9.13. How to choose the best pre-processing <p></p>
<p>9.14. Is pre-processing always useful?</p>
<p>9.15. Real cases</p>
<p>9.16. A library of MATLAB codes</p>
<p>9.17. Exercises</p>
<p>9.18. References</p>
<p> </p>
<p>10. Analytical figures of merit</p>
<p>10.1. Usefulness of figures of merit</p>
<p>10.2. Sensitivity</p>
<p>10.3. Selectivity</p>
<p>10.4. Prediction uncertainty</p>
<p>10.5. Effect of mathematical pre-processing</p>
<p>10.6. Detection limit</p>
<p>10.7. The blank leverage</p>
<p>10.8. Quantitation limit</p>
<p>10.9. Other figures of merit</p>
<p>10.10. A real case</p>
<p>10.11. References</p>
<p> </p>
<p>11. MVC1: software for first-order multivariate calibration</p>
<p>11.1. Downloading and installing the software</p>
11.2. General characteristics<p></p>
<p>11.3. Example 1: determination of bromhexine in anti-cough syrups by UV-visible spectrophotometry</p>
<p>11.4. Example 2: determination of I5 in reaction mixtures by UV-visible spectrophotometry</p>
<p>11.5. Example 3: determination of moisture, fat, protein and starch in corn seeds</p>
<p>11.6. More examples</p>
<p>11.7. Other calibration models</p>
<p>11.8. Free and commercial software</p>
<p>11.9. References</p>
<p> </p>
<p>12. Non-linearity and artificial neural networks</p>
<p>12.1. Linear and non-linear problems</p>
<p>12.2. Artificial neural networks</p>
<p>12.3. Radial basis functions</p>
<p>12.4. Neural networks in MVC1</p>
<p>12.5. A real case</p>
<p>12.6. References</p>
<p>1.2. Univariate and multivariate calibration</p>
<p>1.3. The order and the ways</p>
<p>1.4. Why multivariate calibration?</p>
<p>1.5. Near infrared spectroscopy: the analytical dream</p>
<p>1.6. Multi-way calibration and new advantages</p>
<p>1.7. References</p>
<p> </p>
<p>2. First-order multivariate models: CLS</p>
<p>2.1. Direct and inverse models </p>
<p>2.2. Classical least-squares</p>
<p>2.3. The CLS calibration phase</p>
<p>2.4. Why least-squares? Mathematical requirements</p>
<p>2.5. The CLS prediction phase</p>
<p>2.6. The CLS vector of regression coefficients</p>
<p>2.7. A CLS algorithm</p>
2.8. Validation of the CLS model<p></p>
<p>2.9. Spectral residuals and sample diagnostic</p>
<p>2.10. The first-order advantage</p>
<p>2.11. A real case </p>
<p>2.12. Advantages and limitations of CLS</p>
<p>2.13. Exercises</p>
<p>2.13. References</p>
<p> </p>
<p>3. First-order multivariate models: ILS</p>
<p>3.1. Why calibrating the other way around? A fantastic idea</p>
<p>3.2. The ILS calibration phase</p>
<p>3.3. Mathematical requirements</p>
<p>3.4. The ILS prediction phase</p>
<p>3.5. An ILS algorithm</p>
<p>3.6. The validation of the ILS model</p>
<p>3.7. Advantages and limitations of ILS</p>
<p>3.8. The successive projections algorithm</p>
3.9. A real case<p></p>
<p>3.10. How to improve ILS</p>
<p>3.11. Exercises</p>
<p>3.12. References</p>
<p> </p>
<p>4. Principal component analysis: PCA</p>
4.1. Why compressing the data?<p></p>
<p>4.2. Real and latent variables</p>
<p>4.3. Principal components</p>
<p>4.4. Significant loadings and scores </p>
<p>4.5. Non-significant loadings and scores </p>
<p>4.6. Sample classification with PCA</p>
<p>4.7. Multivariate calibration with PCA</p>
<p>4.8. Exercises</p>
<p>4.9. References</p>
<p> </p>
<p>5. First-order multivariate models: PCR</p>
<p>5.1. Combination of PCA and ILS: another fantastic idea</p>
<p>5.2. Matrix compression and decompression</p>
<p>5.3. The PCR calibration phase</p>
<p>5.4. Mathematical requirements</p>
<p>5.5. The PCR prediction phase</p>
<p>5.6. The PCR vector of regression coefficients</p>
5.7. A PCR algorithm<p></p>
<p>5.8. What is the value of A?</p>
<p>5.9. Advantages and limitations of PCR</p>
<p>5.10. A real case</p>
<p>5.11. What can be better than PCR?</p>
<p>5.12. Exercises</p>
<p>5.13. References</p>
<p> </p>
<p>6. The optimum number of latent variables</p>
<p>6.1. The importance of estimating the optimum A</p>
<p>6.2. Explained variance</p>
<p>6.3. Visual inspection of loadings</p>
<p>6.4. Leave-one-out cross validation</p>
<p>6.5. Cross-validation statistics</p>
<p>6.6. Monte Carlo cross-validation</p>
<p>6.7. Other methods</p>
<p>6.8. The principle of parsimony</p>
6.9. Beyond statistics: physicochemical interpretation of A<p></p>
<p>6.10. Exercises</p>
<p>6.11. References</p>
<p> </p>
<p>7. First-order multivariate models: PLS</p>
<p>7.1. The PLS philosophy</p>
<p>7.2. The PLS calibration phase</p>
<p>7.3. Mathematical requirements</p>
<p>7.4. The number of latent variables</p>
<p>7.5. The PLS prediction phase</p>
<p>7.6. The vector of PLS regression coefficients</p>
7.7. A PLS algorithm<p></p>
<p>7.8. Advantages and limitations of PLS</p>
<p>7.9. A real case</p>
<p>7.10. PLS-1 and PLS-2 models</p>
<p>7.11. Discriminant PLS </p>
<p>7.12. Beyond PLS</p>
7.12. Exercises<p></p>
<p>7.13. References</p>
<p> </p>
<p>8. Comparison of models</p>
<p>8.1. Which is the best model?</p>
<p>8.2. The randomization test</p>
<p>8.3. How the test works</p>
<p>8.4. Algorithm for the randomization test</p>
<p>8.5. PCR, PLS-1 and PLS-2: when, how and why</p>
<p>8.6. Linear and non-linear models: when, how and why</p>
<p>8.7. Tests of multivariate non-linearity</p>
8.8. A real case<p></p>
<p>8.9. Conclusions</p>
<p>8.10. References</p>
<p> </p>
<p>9. Data pre-processing</p>
<p>9.1. Selection of calibration samples</p>
9.2. Calibration outliers <p></p>
<p>9.3. Selection of wavelengths to build the model</p>
<p>9.4. The vector of regression coefficients as selector</p>
<p>9.5. Interval-PLS </p>
<p>9.6. A real case</p>
<p>9.7. Other selection methods</p>
<p>9.8. Mathematical transformation of spectra</p>
<p>9.9. Mean centering</p>
<p>9.10. Smoothing and derivatives: advantages and hazards</p>
<p>9.11. Multiplicative correction</p>
<p>9.12. Other pre-processing methods</p>
9.13. How to choose the best pre-processing <p></p>
<p>9.14. Is pre-processing always useful?</p>
<p>9.15. Real cases</p>
<p>9.16. A library of MATLAB codes</p>
<p>9.17. Exercises</p>
<p>9.18. References</p>
<p> </p>
<p>10. Analytical figures of merit</p>
<p>10.1. Usefulness of figures of merit</p>
<p>10.2. Sensitivity</p>
<p>10.3. Selectivity</p>
<p>10.4. Prediction uncertainty</p>
<p>10.5. Effect of mathematical pre-processing</p>
<p>10.6. Detection limit</p>
<p>10.7. The blank leverage</p>
<p>10.8. Quantitation limit</p>
<p>10.9. Other figures of merit</p>
<p>10.10. A real case</p>
<p>10.11. References</p>
<p> </p>
<p>11. MVC1: software for first-order multivariate calibration</p>
<p>11.1. Downloading and installing the software</p>
11.2. General characteristics<p></p>
<p>11.3. Example 1: determination of bromhexine in anti-cough syrups by UV-visible spectrophotometry</p>
<p>11.4. Example 2: determination of I5 in reaction mixtures by UV-visible spectrophotometry</p>
<p>11.5. Example 3: determination of moisture, fat, protein and starch in corn seeds</p>
<p>11.6. More examples</p>
<p>11.7. Other calibration models</p>
<p>11.8. Free and commercial software</p>
<p>11.9. References</p>
<p> </p>
<p>12. Non-linearity and artificial neural networks</p>
<p>12.1. Linear and non-linear problems</p>
<p>12.2. Artificial neural networks</p>
<p>12.3. Radial basis functions</p>
<p>12.4. Neural networks in MVC1</p>
<p>12.5. A real case</p>
<p>12.6. References</p>
Rubrieken
- advisering
- algemeen management
- coaching en trainen
- communicatie en media
- economie
- financieel management
- inkoop en logistiek
- internet en social media
- it-management / ict
- juridisch
- leiderschap
- marketing
- mens en maatschappij
- non-profit
- ondernemen
- organisatiekunde
- personal finance
- personeelsmanagement
- persoonlijke effectiviteit
- projectmanagement
- psychologie
- reclame en verkoop
- strategisch management
- verandermanagement
- werk en loopbaan