Statistics Toolbox

Statistics Toolbox Overview   3:06  

Get an overview of Statistics Toolbox capabilities.

An Introduction to Dataset Arrays   5:31  

Improve productivity with statistical arrays.

An Introduction to Joins   5:01  

Combine fields from two dataset arrays using a variable that is present in both.

An Introduction to Classification   9:04

Develop predictive models for classifying data.

Simulating Dependent Random Variables Using Copulas  

Create distributions that model correlated multivariate data.

Selecting Features for Classifying High-Dimensional Data  

Use sequential feature selection and filtering to reduce model complexity.

Partial Least Squares Regression and Principal Components Regression  

Model a response variable in the presence of highly correlated predictors.

Fitting Data with Generalized Linear Models  

Fit regression models where the error terms are not normally distributed (logistic regression, Poisson regression).

Nonparametric Fitting   4:08

Develop a predictive model without specifying a function that describes the relationship between variables.

Fitting Data with Generalized Linear Models  

Fit regression models where the error terms are not normally distributed (logistic regression, Poisson regression).

Partial Least Squares Regression and Principal Components Regression  

Model a response variable in the presence of highly correlated predictors.

Pitfalls in Fitting Nonlinear Models by Transforming to Linearity  

Transform nonlinear regression problems into linear regression problems.

Weighted Nonlinear Regression  

Fit a nonlinear regression when the error variance is not constant.

Fitting an Orthogonal Regression Using Principal Components Analysis  

Implement Deming regression (total least squares).

Time Series Regression of Airline Passenger Data  

Model data that exhibits seasonality and autocorrelation.

An Introduction to Classification   9:04

Develop predictive models for classifying data.

Classification  

Classify data into known groups using Naïve Bayes classifier and linear/quadratic discriminant analysis.

Selecting Features for Classifying High-Dimensional Data  

Use sequential feature selection and filtering to reduce model complexity.

Cluster Analysis  

Use k-means and hierarchical clustering to discover natural groupings in data.

Working with Probability Distributions   8:15

Fit distributions to empirical data, and visually explore the effects of changing parameters on the shape of a distribution.

Simulating Dependent Random Variables Using Copulas  

Create distributions that model correlated multivariate data.

Analyzing Survival or Reliability Data  

Model the time to failure of a piece of equipment or the survival time of an organism.

Fitting Custom Univariate Distributions, Part 1  

Perform maximum likelihood estimation on truncated, weighted, or bimodal data.

Fitting Custom Univariate Distributions, Part 2  

Overcome numerical problems in performing maximum likelihood estimation on censored data.

Modeling Data with the Generalized Extreme Value Distribution  

Model extreme, rare events (equity risks, earthquakes, hundred-year floods).

Modeling Tail Data with the Generalized Pareto Distribution  

Model low-density tails of distributions.

Curve Fitting and Distribution Fitting  

Determine when to apply distribution fitting techniques instead of using curve fitting.

Fitting a Univariate Distribution Using Cumulative Probabilities  

Use cumulative probabilities to fit a univariate distribution when MLE is infeasible (for example, models with threshold parameters).

Nonparametric Estimates of Cumulative Distribution Functions and Their Inverses  

Compute a non- or semi-parametric CDF or inverse CDF from data.

Selecting Features for Classifying High-Dimensional Data  

Use sequential feature selection and filtering to reduce model complexity.

Visualizing Multivariate Data  

Visualize multivariate data using scatter plot matrices, parallel coordinate plots, and other specialized graphs.

Partial Least Squares Regression and Principal Components Regression  

Model a response variable in the presence of highly correlated predictors.

Factor Analysis  

Fit a model to multivariate data with correlated variables using a reduced set of latent factors.

Classical Multidimensional Scaling  

Assign points to spatial locations to approximate the known distances between them.

Non-Classical Multidimensional Scaling  

Assign points to spatial locations to mimic known similarity or dissimilarity measures.

Selecting a Sample Size  

Calculate the sample size required for a hypothesis test.

Improving an Engine Cooling Fan Using Design for Six Sigma Techniques  

Apply Design of Experiments and Statistical Process Control to a Six Sigma analysis problem.