David Booth - Profile on Academia.edu (original) (raw)
Papers by David Booth
Data Analysis Using Regression Models: The Business Perspective
The American Statistician, May 1, 1997
Time Series (3rd ed.)
Technometrics, Mar 12, 2012
Data Analysis Using Regression Models: The Business Perspective
The American Statistician, 1997
Analysis of Incomplete Multivariate Data
Technometrics, May 1, 2000
Chapter 5 moves the focus to object detection. Here, the model is developed using transformed spa... more Chapter 5 moves the focus to object detection. Here, the model is developed using transformed space-based techniques. The authors mention the use of moment-invariant techniques and provide a few references. There is also some illustration of cases in which these techniques would work well. The book does not even mention what these techniques are, however. It is difficult, therefore, to appreciate what the supposed merits or drawbacks are. For the transformed space-based techniques, it is not clear how one should choose the transformation. The methodology is illustrated on some test examples. It is not clear what advantage these methods would have over other methodology in use [such as the use of object processes; see Baddeley and van Lieshout (1993)] in the statistics and imaging literature, especially because the objects are very easy to detect in each of the supplied examples. Once again, there is no mention of these other methods. Chapter 6 develops image segmentation using a nested design algorithm. The authors advocate a top-down splitting and merging strategy to segmentation, starting from the top left corner of the image. The methodology is illustrated on the “Lena” image, but though the segmented image seems to look similar to the original, we wonder whether the results are really any good. For example, we could detect at least five segments in the face of “Lena” (Fig. 6.12) and six on her hat, but should not a good algorithm detect these in the same (and only one) segment for each? Once again, there is no discussion at all on an interpretation of these results and the drawbacks and no comparison with other procedures (e.g., see Chen, Lin, and Chen 1991). Chapter 9 addresses the aspect of image restoration. Two restoration procedures based on a two-stage approach are presented. In this, edge detectors based on generalized least squares and Youden square masks are presented. The examples here work well, though once again, it is not readily apparent how the different 5 x 5 masks are merged. Other approaches for image restoration (see Amit, Grenander, and Piccioni 1991) are not even mentioned. The authors are also quite sloppy with the notations and definitions. Here are just a few examples. On page 14 [e.g., in Eq. (2.28)], the degrees of freedom for the denominator should be m(n l), not m2 m. In Equation (2.58), +y..k should be -y,,k. For the Tukey method on page 30, the confidence interval is & i Ts( &), not just Ts( &). In Equation (3.26) on page 49, s2 cannot be equal to 0’; s2 is only an estimator of c2. In Equation (3.42) on page 53, the error sum of squares cannot depend on ,!?; it should involve only the estimator of p. On the bottom of page 69, it is mentioned, “Based on numerous simulations on test images, 2isselectedastheadditionalteststatistic.”Whatdoesthismean?2 is selected as the additional test statistic.” What does this mean? 2isselectedastheadditionalteststatistic.”Whatdoesthismean?2 cannot be a test statistic, being an unknown parameter; see (4.5). In Equation (6.3) on page 134, -13~~~1 should be +03c2). What is A, in Equation (6.7)? Below Equation (A.l) on page 145, it is mentioned that A, = c, q3. It is not clear if this is the right quantity in Equation (6.7). In Equation (6.8) +yZ., should be -yi.. On page 148, below Equation (7.8) Qa should not involve /3. We find the discussion in Section 8.1, on Bayes and maximum likelihood estimation, very vague. The authors talk of p(a), p(r/a), and E(r/a), without explaining what these quantities are. In short, the book should not be consulted for learning anything about ANOVA or image analysis. In writing this book, the authors were hoping that “this book will stimulate a wider acceptance of the methodology of image processing suggested here” (see Preface, p. xii). We have serious doubts about the book achieving this goal.
Time Series, Forecasting, Simulation, Applications
Technometrics, May 1, 1994
Book Reviews: Technometrics
Technometrics, Oct 1, 2018
This section reviews those books whose content and level reflect the general editorial policy of ... more This section reviews those books whose content and level reflect the general editorial policy of Technometrics...
The Cross-Entropy Method, by Reuven Y. Rubinstein and Dirk P. Kroese
Technometrics, 2008
Time Series Analysis
Technometrics, 1997
Ceres Plots
Encyclopedia of Statistical Sciences, Aug 15, 2006
Statistical Methods for SPC and TQM
Technometrics, Nov 1, 1994
Introduction - statistics, SPC and total quality data collection and graphical summaries numerica... more Introduction - statistics, SPC and total quality data collection and graphical summaries numerical data summaries - location and dispersion probability and distribution sampling, estimation and confidence simple tests of hypotheses - "significance tests" control charts for process management and improvement control charts for average and variation control charts for "single-valued" observations control charts for attributes and events control charts - problems and special cases cusum methods process capability - attributes, events and normally distributed data capability - non-normal distributions evaluating the precision of a measurement system (gauge capability) getting more from control chart data SPC in "non-product" applications. Appendices: linear combinations of independent variables linear combination of correlated variables products and quotients non-linear functions dealing with composite functions other useful standard errors.
Journal of Chemical Information and Computer Sciences, Mar 1, 1997
It has previously been shown that smoothing algorithms can provide the basis for a method to dete... more It has previously been shown that smoothing algorithms can provide the basis for a method to detect nuclear material diversions and losses and moreover can also provide a general approach to industrial statistical process control. The present paper extends this result by showing that a set of robust smoothers also produces equivalent methods that can be used in nuclear material safeguards algorithms. Further, it is shown that these smoothers are somewhat more sensitive to loss points than the previously studied smoothers. The method is illustrated on real data.
Journal of Chemical Information and Computer Sciences, Mar 1, 1997
It has previously been shown that smoothing algorithms can provide the basis for methods to detec... more It has previously been shown that smoothing algorithms can provide the basis for methods to detect nuclear material losses and moreover can also provide a general approach to industrial statistical process control. The present paper extends this result by showing that a set of robust smoothers also produces methods that can be used in statistical process control. Further, it is shown that these smoothers are somewhat more sensitive to out of control points than those methods previously studied. The methods are successfully illustrated on chemical process data.
A method for early discovery of poisoning in catalytic chemical processes
Journal of Chemical Information and Computer Sciences, May 1, 1985
Computers and Biomedical Research, Feb 1, 1986
Classification is one of the fundamental goals of science and is basic to the diagnosis of diseas... more Classification is one of the fundamental goals of science and is basic to the diagnosis of disease. Unfortunately, classifying objects (e.g., patients) on the basis of clinical and/or laboratory experimental observations into various groups can be difficult when the groups overlap or contain outlying points. Recently, Broffitt, Randles, and co-workers proposed a procedure, robust partial discriminant analysis (RPDA) for dealing with such problems, but testing of the procedure was limited to Monte Carlo simulation. In this study, RPDA was applied to real data, in order to compare its effectiveness with ordinary discriminant analysis, as well as to determine if RPDA was a suitable procedure to use to classify chemical compounds on the basis of experimental observations and as a tool in the diagnosis of disease (in particular, multiple sclerosis and thyrotoxicosis), with data based on experimental and clinical observations. The resulting RPDA classifications were an improvement over those obtained from ordinary discriminant analysis.
Technometrics, Nov 1, 1999
A fractal dimension-based method for statistical process control
International Journal of Operational Research, 2012
... Booth, DE, Zhu, DX, Baker, DL and Hamburg, JH (2005) 'Recent chemometric approaches ... V... more ... Booth, DE, Zhu, DX, Baker, DL and Hamburg, JH (2005) 'Recent chemometric approaches ... V. (2004) 'Wavelet-based multiscale statistical process monitoring: a literature review', IEEE Transactions ... Grant, EL and Leavenworth, RS (1980) Statistical Quality Control (5th ed.). New ...
In this paper, we present experimental results of fuzzy clustering and two self-organizing neural... more In this paper, we present experimental results of fuzzy clustering and two self-organizing neural networks used as classification tools for identifying potentially failing banks. We first describe the distinctive characteristics of fuzzy clustering algorithm, which provides probability of the likelihood of bank failure. We then perform the comparison between the results of the closest hard partitioning of fuzzy clustering and of two self-organizing neural networks and present our results as the ranking structure of relative bankruptcy likelihood. Our findings indicate that both the fuzzy clustering and self-organizing neural networks are promising classification tools for identifying potentially failing banks.
Chemosensory Perception, 2011
Among the early indications of the existence of supertasters, tasters and non-tasters was a trimo... more Among the early indications of the existence of supertasters, tasters and non-tasters was a trimodal distribution of sensitivities to the taste of low concentrations of caffeine in water. A similar three peaks of prevalence have now been seen in the concentration of caffeine in the individual's usual coffee drink that is perceived as the most preferred, measured by normed multiple discrimination psychophysics. Furthermore, the mode at the lowest concentrations of caffeine in coffee-the putative supertasters-was clearer for preference than for bitterness. This was because rated preference for a sample of coffee as a whole can be more sensitive to differences in level of caffeine than rating specifically how bitter it is. This criterion of gustatory performance is independent of the particular scores given by an assessor or the subjective experiencing of any sensations. Hence perception of the tastant in a familiar context is capable of picking out supertasters and non-tasters from the tasters.
Data Analysis Using Regression Models: The Business Perspective
The American Statistician, May 1, 1997
Time Series (3rd ed.)
Technometrics, Mar 12, 2012
Data Analysis Using Regression Models: The Business Perspective
The American Statistician, 1997
Analysis of Incomplete Multivariate Data
Technometrics, May 1, 2000
Chapter 5 moves the focus to object detection. Here, the model is developed using transformed spa... more Chapter 5 moves the focus to object detection. Here, the model is developed using transformed space-based techniques. The authors mention the use of moment-invariant techniques and provide a few references. There is also some illustration of cases in which these techniques would work well. The book does not even mention what these techniques are, however. It is difficult, therefore, to appreciate what the supposed merits or drawbacks are. For the transformed space-based techniques, it is not clear how one should choose the transformation. The methodology is illustrated on some test examples. It is not clear what advantage these methods would have over other methodology in use [such as the use of object processes; see Baddeley and van Lieshout (1993)] in the statistics and imaging literature, especially because the objects are very easy to detect in each of the supplied examples. Once again, there is no mention of these other methods. Chapter 6 develops image segmentation using a nested design algorithm. The authors advocate a top-down splitting and merging strategy to segmentation, starting from the top left corner of the image. The methodology is illustrated on the “Lena” image, but though the segmented image seems to look similar to the original, we wonder whether the results are really any good. For example, we could detect at least five segments in the face of “Lena” (Fig. 6.12) and six on her hat, but should not a good algorithm detect these in the same (and only one) segment for each? Once again, there is no discussion at all on an interpretation of these results and the drawbacks and no comparison with other procedures (e.g., see Chen, Lin, and Chen 1991). Chapter 9 addresses the aspect of image restoration. Two restoration procedures based on a two-stage approach are presented. In this, edge detectors based on generalized least squares and Youden square masks are presented. The examples here work well, though once again, it is not readily apparent how the different 5 x 5 masks are merged. Other approaches for image restoration (see Amit, Grenander, and Piccioni 1991) are not even mentioned. The authors are also quite sloppy with the notations and definitions. Here are just a few examples. On page 14 [e.g., in Eq. (2.28)], the degrees of freedom for the denominator should be m(n l), not m2 m. In Equation (2.58), +y..k should be -y,,k. For the Tukey method on page 30, the confidence interval is & i Ts( &), not just Ts( &). In Equation (3.26) on page 49, s2 cannot be equal to 0’; s2 is only an estimator of c2. In Equation (3.42) on page 53, the error sum of squares cannot depend on ,!?; it should involve only the estimator of p. On the bottom of page 69, it is mentioned, “Based on numerous simulations on test images, 2isselectedastheadditionalteststatistic.”Whatdoesthismean?2 is selected as the additional test statistic.” What does this mean? 2isselectedastheadditionalteststatistic.”Whatdoesthismean?2 cannot be a test statistic, being an unknown parameter; see (4.5). In Equation (6.3) on page 134, -13~~~1 should be +03c2). What is A, in Equation (6.7)? Below Equation (A.l) on page 145, it is mentioned that A, = c, q3. It is not clear if this is the right quantity in Equation (6.7). In Equation (6.8) +yZ., should be -yi.. On page 148, below Equation (7.8) Qa should not involve /3. We find the discussion in Section 8.1, on Bayes and maximum likelihood estimation, very vague. The authors talk of p(a), p(r/a), and E(r/a), without explaining what these quantities are. In short, the book should not be consulted for learning anything about ANOVA or image analysis. In writing this book, the authors were hoping that “this book will stimulate a wider acceptance of the methodology of image processing suggested here” (see Preface, p. xii). We have serious doubts about the book achieving this goal.
Time Series, Forecasting, Simulation, Applications
Technometrics, May 1, 1994
Book Reviews: Technometrics
Technometrics, Oct 1, 2018
This section reviews those books whose content and level reflect the general editorial policy of ... more This section reviews those books whose content and level reflect the general editorial policy of Technometrics...
The Cross-Entropy Method, by Reuven Y. Rubinstein and Dirk P. Kroese
Technometrics, 2008
Time Series Analysis
Technometrics, 1997
Ceres Plots
Encyclopedia of Statistical Sciences, Aug 15, 2006
Statistical Methods for SPC and TQM
Technometrics, Nov 1, 1994
Introduction - statistics, SPC and total quality data collection and graphical summaries numerica... more Introduction - statistics, SPC and total quality data collection and graphical summaries numerical data summaries - location and dispersion probability and distribution sampling, estimation and confidence simple tests of hypotheses - "significance tests" control charts for process management and improvement control charts for average and variation control charts for "single-valued" observations control charts for attributes and events control charts - problems and special cases cusum methods process capability - attributes, events and normally distributed data capability - non-normal distributions evaluating the precision of a measurement system (gauge capability) getting more from control chart data SPC in "non-product" applications. Appendices: linear combinations of independent variables linear combination of correlated variables products and quotients non-linear functions dealing with composite functions other useful standard errors.
Journal of Chemical Information and Computer Sciences, Mar 1, 1997
It has previously been shown that smoothing algorithms can provide the basis for a method to dete... more It has previously been shown that smoothing algorithms can provide the basis for a method to detect nuclear material diversions and losses and moreover can also provide a general approach to industrial statistical process control. The present paper extends this result by showing that a set of robust smoothers also produces equivalent methods that can be used in nuclear material safeguards algorithms. Further, it is shown that these smoothers are somewhat more sensitive to loss points than the previously studied smoothers. The method is illustrated on real data.
Journal of Chemical Information and Computer Sciences, Mar 1, 1997
It has previously been shown that smoothing algorithms can provide the basis for methods to detec... more It has previously been shown that smoothing algorithms can provide the basis for methods to detect nuclear material losses and moreover can also provide a general approach to industrial statistical process control. The present paper extends this result by showing that a set of robust smoothers also produces methods that can be used in statistical process control. Further, it is shown that these smoothers are somewhat more sensitive to out of control points than those methods previously studied. The methods are successfully illustrated on chemical process data.
A method for early discovery of poisoning in catalytic chemical processes
Journal of Chemical Information and Computer Sciences, May 1, 1985
Computers and Biomedical Research, Feb 1, 1986
Classification is one of the fundamental goals of science and is basic to the diagnosis of diseas... more Classification is one of the fundamental goals of science and is basic to the diagnosis of disease. Unfortunately, classifying objects (e.g., patients) on the basis of clinical and/or laboratory experimental observations into various groups can be difficult when the groups overlap or contain outlying points. Recently, Broffitt, Randles, and co-workers proposed a procedure, robust partial discriminant analysis (RPDA) for dealing with such problems, but testing of the procedure was limited to Monte Carlo simulation. In this study, RPDA was applied to real data, in order to compare its effectiveness with ordinary discriminant analysis, as well as to determine if RPDA was a suitable procedure to use to classify chemical compounds on the basis of experimental observations and as a tool in the diagnosis of disease (in particular, multiple sclerosis and thyrotoxicosis), with data based on experimental and clinical observations. The resulting RPDA classifications were an improvement over those obtained from ordinary discriminant analysis.
Technometrics, Nov 1, 1999
A fractal dimension-based method for statistical process control
International Journal of Operational Research, 2012
... Booth, DE, Zhu, DX, Baker, DL and Hamburg, JH (2005) 'Recent chemometric approaches ... V... more ... Booth, DE, Zhu, DX, Baker, DL and Hamburg, JH (2005) 'Recent chemometric approaches ... V. (2004) 'Wavelet-based multiscale statistical process monitoring: a literature review', IEEE Transactions ... Grant, EL and Leavenworth, RS (1980) Statistical Quality Control (5th ed.). New ...
In this paper, we present experimental results of fuzzy clustering and two self-organizing neural... more In this paper, we present experimental results of fuzzy clustering and two self-organizing neural networks used as classification tools for identifying potentially failing banks. We first describe the distinctive characteristics of fuzzy clustering algorithm, which provides probability of the likelihood of bank failure. We then perform the comparison between the results of the closest hard partitioning of fuzzy clustering and of two self-organizing neural networks and present our results as the ranking structure of relative bankruptcy likelihood. Our findings indicate that both the fuzzy clustering and self-organizing neural networks are promising classification tools for identifying potentially failing banks.
Chemosensory Perception, 2011
Among the early indications of the existence of supertasters, tasters and non-tasters was a trimo... more Among the early indications of the existence of supertasters, tasters and non-tasters was a trimodal distribution of sensitivities to the taste of low concentrations of caffeine in water. A similar three peaks of prevalence have now been seen in the concentration of caffeine in the individual's usual coffee drink that is perceived as the most preferred, measured by normed multiple discrimination psychophysics. Furthermore, the mode at the lowest concentrations of caffeine in coffee-the putative supertasters-was clearer for preference than for bitterness. This was because rated preference for a sample of coffee as a whole can be more sensitive to differences in level of caffeine than rating specifically how bitter it is. This criterion of gustatory performance is independent of the particular scores given by an assessor or the subjective experiencing of any sensations. Hence perception of the tastant in a familiar context is capable of picking out supertasters and non-tasters from the tasters.