Linked e-resources
Details
Table of Contents
Intro; Contents; 1 Introduction; References; 2 Drug Use and Personality Profiles; 2.1 Definitions of Drugs and Drug Usage; 2.2 Personality Traits; 2.3 How Many Inputs Do the Predictive Models Have: 5, 30, 60, or 240?; 2.4 The Problem of the Relationship Between Personality Traits and Drug Consumption; 2.5 New Data Set Open for Use; 2.6 First Results in Brief; References; 3 Methods of Data Analysis; 3.1 Preprocessing; 3.1.1 Descriptive Statistics: Mean, Variance, Covariance, Correlation, Information Gain; 3.1.2 Input Feature Normalisation; 3.2 Input Feature Transformation
3.2.1 Principal Component Analysis-PCA3.2.2 Quantification of Categorical Input Variables; 3.2.3 Input Feature Ranking; 3.3 Classification and Risk Evaluation; 3.3.1 Single Attribute Predictors; 3.3.2 Criterion for Selecting the Best Method; 3.3.3 Linear Discriminant Analysis (LDA); 3.3.4 Logistic Regression (LR); 3.3.5 k-Nearest Neighbours (kNN); 3.3.6 Decision Tree (DT); 3.3.7 Random Forest (RF); 3.3.8 Gaussian Mixture (GM); 3.3.9 Probability Density Function Estimation (PDFE); 3.3.10 Naïve Bayes (NB); 3.4 Visualisation on the Nonlinear PC Canvas: Elastic Maps; References
4 Results of Data Analysis4.1 Descriptive Statistics and Psychological Profile of Illicit Drug Users; 4.2 Distribution of Number of Drugs Used; 4.3 Sample Mean and Population Norm; 4.4 Deviation of the Groups of Drug Users from the Sample Mean; 4.5 Significant Differences Between Groups of Drug Users and Non-users; 4.6 Correlation Between Usage of Different Drugs; 4.7 Ranking of Input Features; 4.8 Selection of the Best Classifiers for the Decade-Based Classification Problem; 4.9 Correlation Pleiades of Drugs; 4.10 Overoptimism Problem
4.11 User/Non-user Classification by Linear Discriminant for Ecstasy and Heroin4.12 Separation of Heroin Users from Ecstasy Users: What Is the Difference?; 4.13 Significant Difference Between Benzodiazepines, Ecstasy, and Heroin; 4.14 A Tree of Linear Discriminants: No Essential Improvements; 4.15 Visualisation on Nonlinear PCA Canvas; 4.16 Risk Maps; References; 5 Summary; References; 6 Discussion; References; A Main Tables; A.1 Psychological Profiles of Drug Users and Non-users; A.2 Correlation Between Consumption of Different Drugs; A.3 Linear Discriminants for User/Non-user Separation
3.2.1 Principal Component Analysis-PCA3.2.2 Quantification of Categorical Input Variables; 3.2.3 Input Feature Ranking; 3.3 Classification and Risk Evaluation; 3.3.1 Single Attribute Predictors; 3.3.2 Criterion for Selecting the Best Method; 3.3.3 Linear Discriminant Analysis (LDA); 3.3.4 Logistic Regression (LR); 3.3.5 k-Nearest Neighbours (kNN); 3.3.6 Decision Tree (DT); 3.3.7 Random Forest (RF); 3.3.8 Gaussian Mixture (GM); 3.3.9 Probability Density Function Estimation (PDFE); 3.3.10 Naïve Bayes (NB); 3.4 Visualisation on the Nonlinear PC Canvas: Elastic Maps; References
4 Results of Data Analysis4.1 Descriptive Statistics and Psychological Profile of Illicit Drug Users; 4.2 Distribution of Number of Drugs Used; 4.3 Sample Mean and Population Norm; 4.4 Deviation of the Groups of Drug Users from the Sample Mean; 4.5 Significant Differences Between Groups of Drug Users and Non-users; 4.6 Correlation Between Usage of Different Drugs; 4.7 Ranking of Input Features; 4.8 Selection of the Best Classifiers for the Decade-Based Classification Problem; 4.9 Correlation Pleiades of Drugs; 4.10 Overoptimism Problem
4.11 User/Non-user Classification by Linear Discriminant for Ecstasy and Heroin4.12 Separation of Heroin Users from Ecstasy Users: What Is the Difference?; 4.13 Significant Difference Between Benzodiazepines, Ecstasy, and Heroin; 4.14 A Tree of Linear Discriminants: No Essential Improvements; 4.15 Visualisation on Nonlinear PCA Canvas; 4.16 Risk Maps; References; 5 Summary; References; 6 Discussion; References; A Main Tables; A.1 Psychological Profiles of Drug Users and Non-users; A.2 Correlation Between Consumption of Different Drugs; A.3 Linear Discriminants for User/Non-user Separation