OLytics Ontology

Ontology for anaLytics
Top Twenty-Five Techniques

Techniques: Prediction

Technique Description ScalaTion Class Status
Simple Linear Regression Fit parameters b0 and b1 in y = b0 + b1 x SimpleRegression implemented
Multiple Linear Regression Fit parameter vector b in y = b dot x Regression implemented
Polynomial Regression Fit parameter vector b in y = b dot x where x is formed from powers of parameter t PolyRegression implemented
Nonlinear Regression Fit parameter vector b in y = f(b, x) when f is nonlinear in b NonLinRegression implemented
Perceptron Adjust weights so the output scalar y is predicted from the input vector x Perceptron implemented
Neural Networks Adjust weights so the output vector y is predicted from the input vector x NeuralNet under development
Radial Basis Functions . . .
ARMA Time Series Analysis Auto-Regressive, Moving-Average models ARMA under development
ARIMA Time Series Analysis . . .
Principal Component Analysis Project a data matrix along its main eigenvectors to reduce dimensionality PrincipalComponents implemented
Factor Analysis Analyze a data matrix in terms of given set of factors. FactorAnalysis started
Canonical Correlation Analysis Analyze/maximize the correlation between two data matrices. CanCorrelation started

Techniques: Classification

Technique Description ScalaTion Class Status
kNN Classifier Classify based on k Nearest Neighbors KNN_Classifier implemented
Naive Bayes Integer Classifier Classify integer data based on product of conditional probabilities P(X_i | C) NaiveBayesInt implemented
Naive Bayes Classifier Classify real data based on product of conditional probabilities P(X_i | C) NaiveBayes implemented
Bayesian Network Classify data based on product of conditional probabilities P(X_i | C, {X_j}) BayesNetwork under development
ID3 Decision Trees Classify data using simple decision trees DecisionTreeID3 implemented
C45 Decision Trees Classify data using decision trees that handle continuous values DecisionTreeC45 implemented
Support Vector Machines Classify data by finding optimal separating hyperplane SupportVectorMachine implemented
Logistic Regression Used when the dependent variable y is binary LogisticRegression implemented
CART Classification And Regression Trees . .

Techniques: Clustering

Technique Description ScalaTion Class Status
K-Means Clustering Cluster the data into k clusters KMeansClustering implemented
Hierachical Clustering Cluster the data into hierachical groups HierClustering implemented
Markov Clustering Cluster the nodes in a graph MarkovClustering implemented
Mixture Model . . .

Taxonomy

Ontology