Java C R
Clone or download

README.md

rapaio

Statistics, data mining and machine learning toolbox in Java.

build status codecov.io codecov.io

codecov.io

Rapaio Manual is published on gitbooks.com. Using the previous link you can select the format of manual. To read it directly online you can use this direct link: Rapaio Manual - read on-line.

NEW: The manual contains a Tutorial on Kaggle's Titanic Competition!

Acknowledgements

Many thanks to ej-technologies GmbH for providing an open source license for their Java Profiler and to JetBrains for providing an open source license for their Java IDE.

Stable Features

For each feature there are some notes regarding the development stage. If there are no notes it means the feature is considered to be fully implemented and well tested.

Core

  • Special Math functions
  • Maximum, Minimum, Mode (only for nominal values), Sum, Mean, Variance, Quantiles
  • Online Statistics: minimum, maximum, count, mean, variance, standard deviation, skewness, kurtosis

Core tools

  • DVector
  • DTable
  • Distance Matrix

Correlations

  • Pearson product-moment coefficient
  • Spearman's rank correlation coefficient

Distributions

  • Bernoulli
  • Binomial
  • Poisson
  • Normal/Gaussian
  • Student t
  • ChiSquare
  • Discrete Uniform
  • Continuous Uniform
  • Hypergeometric
  • Gamma
  • Empirical KDE (gaussian, epanechnikov, cosine, tricube, biweight, triweight, triangular, uniform)

Sampling

  • SamplingTool
    • generates discrete integer samples with/without replacement, weighted/non-weighted
    • offers utility methods for bootstraps, simple random, stratified sampling
  • Samplers used in machine learning algorithms

Hypothesis Testing

  • z test
    • one sample test for testing the sample mean
    • two unpaired samples test for testing difference of the sample means
    • two paired samples test for testing sample mean of the differences
  • t test
    • one sample test for testing the sample mean
    • two unpaired samples t test with same variance
    • two unpaired samples Welch t test with different variances
    • two paired samples test for testing sample mean of differences
  • Kolmogorov Smirnoff KS test
    • one sample test for testing if a sample belongs to a distribution
    • two samples test for testing if both samples comes from the same distribution
  • Pearson Chi-Square tests
    • goodness of fit
    • independence test
    • conditional independence test
  • Anderson-Darling goodness of fit
    • normality test

Frame Filters

  • FFJitter - add jitter to data according with a noise distribution
  • FFAddIntercept - add an intercept variable to a given data set
  • FFMapVars - select some variables according with a VRange pattern
  • FFRemoveVars - removes some variables according with a VRange pattern
  • FFStandardize - standardize variables from a given data frame
  • FFRandomProjection - project a data frame onto random projections

Var filters

  • VFCumulativeSum - builds a numeric vector with a cumulative sum
  • VFJitter - adds noise to a given numeric vector according with a noise distribution
  • VFRefSort - sorts a variable according with a given set of row comparators
  • VFShuffle - shuffles values from a variable
  • VFSort - sorts a variable according with default comparator
  • VFStandardize - standardize values from a given numeric variable
  • VFToIndex - transforms a variable into an index type using a lambda
  • VFToNumeric - transforms a variable into numeric using a lambda
  • VFTransformPower - transform a variable withg power transform
  • VFUpdate - updates a variable using a lambda on VSpot
  • VFUpdateIndex - updates a variable using a lambda on index value
  • VFUpdateLabel - updates a variable using a lambda on label value
  • VFUpdateValue - updates a variable using a lambda on double value

Evaluation

  • Confusion Matrix
  • Receiver Operator Characteristic - ROC curves and ROC Area
  • Root Mean Square Error
  • Mean Absolute Error
  • Gini / Normalized Gini

Analysis

  • Fast Fourier Transform
  • Principal Components Analysis
  • Fischer Linear Discriminant Analysis

Classification

  • Bayesian: NaiveBayes (GaussianPdf, EmpiricalPdf, MultinomialPmf)
  • Linear: BinaryLogistic
  • Rule: OneRule
  • Decision Trees - CTree: DecisionStump, ID3, C45, CART
    • purity: entropy, infogain, gain ration, gini index
    • weight on instances
    • split: numeric binary, nominal binary, nominal full
    • missing value handling: ignore, random, majority, weighted
    • reduced-error pruning
    • variable importance: frequency, gain and permutation based
  • Ensemble: CForest - Bagging, Random Forests
  • Boosting: AdaBoost.SAMME
  • SVM: BinarySMO (Platt)

Regression

  • Simple: ConstantRegression
  • Simple: L1Regression
  • Simple: L2Regression
  • Simple: RandomValueRegressor
  • Tree: CART (no pruning)
  • Tree: C45 (no pruning)
  • Tree: DecisionStump
  • LinearRegression (multiple targets, only numerical attributes)

Clustering

  • Cluster Silhouette
  • KMeans clustering
  • Minkowski Weighted KMeans

Time Series

  • Acf (correlation, covariance)
  • Pacf

Graphics

  • QQ Plot
  • Box Plot
  • Histogram
  • 2d Histogram
  • Plot function line
  • Plot vertical/horizontal/ab line
  • Plot lines
  • Plot points
  • Density line KDE
  • ROC Curve
  • Discrete Vertical Lines
  • Segment2D

Matrices and vectors

  • Numeric vector operations
  • Basic matrix operations and matrix decompositions

Experminental Stage Features

Classification

  • Boosting: GBT (Gradient Boosting Trees) Classifier
  • Ensemble: SplitClassifier

Regression

  • Boost: GBT (Gradient Boosting Tree) Regressor
  • NNet: MultiLayer Perceptron Regressor

Graphics

All the graphics components are in usable state. However the graphics customization needs further improvements in order to make the utilization easier.

  • Plot legend
  • BarChart