Package ML

module containing machine learning code

Submodules

[hide private]

ML.AnalyzeComposite: command line utility to report on the contributions of descriptors to tree-based composite models Usage: AnalyzeComposite [optional args] <models> <models>: file name(s) of pickled composite model(s) (this is the name of the db table if using a database) Optional Arguments: -n number: the number of levels of each model to consider -d dbname: the database from which to read the models -N Note: the note string to search for to pull models from the database -X: Send the results to Excel.
ML.BuildComposite: command line utility for building composite models #DOC **Usage** BuildComposite [optional args] filename Unless indicated otherwise (via command line arguments), _filename_ is a QDAT file.
ML.Cluster
- ML.Cluster.Butina: Implementation of the clustering algorithm published in:...
- ML.Cluster.ClusterUtils: utility functions for clustering
- ML.Cluster.ClusterVis: Cluster tree visualization using Sping
- ML.Cluster.Clustering
- ML.Cluster.Clusters: contains the Cluster class for representing hierarchical cluster trees
- ML.Cluster.Murtagh: Interface to the C++ Murtagh hierarchic clustering code
- ML.Cluster.Resemblance: code for dealing with resemblance (metric) matrices...
- ML.Cluster.Standardize: contains code for standardization of data matrices for clustering
ML.Composite
- ML.Composite.AdjustComposite: functionality to allow adjusting composite model contents
- ML.Composite.BayesComposite: code for dealing with Bayesian composite models For a model to be useable here, it should support the following API: - _ClassifyExample(example)_, returns a classification Other compatibility notes: 1) To use _Composite.Grow_ there must be some kind of builder functionality which returns a 2-tuple containing (model,percent accuracy).
- ML.Composite.Composite: code for dealing with composite models For a model to be useable here, it should support the following API: - _ClassifyExample(example)_, returns a classification Other compatibility notes: 1) To use _Composite.Grow_ there must be some kind of builder functionality which returns a 2-tuple containing (model,percent accuracy).
ML.CompositeRun: contains a class to store parameters for and results from...
ML.Data
- ML.Data.DataUtils: Utilities for data manipulation **FILE FORMATS:** - *.qdat files* contain quantized data suitable for feeding to learning algorithms.
- ML.Data.FindQuantBounds
- ML.Data.MLData: classes to be used to help work with data sets
- ML.Data.Quantize: Automatic search for quantization bounds This uses the expected informational gain to determine where quantization bounds should lie.
- ML.Data.SplitData
- ML.Data.Stats: various statistical operations on data
- ML.Data.Transforms
ML.DecTree: Here we're implementing the Decision Tree stuff found in Chapter 3 of Tom Mitchell's Machine Learning Book.
- ML.DecTree.BuildQuantTree:
- ML.DecTree.BuildSigTree:
- ML.DecTree.CrossValidate: handles doing cross validation with decision trees This is, perhaps, a little misleading.
- ML.DecTree.DecTree: Defines the class _DecTreeNode_, used to represent decision trees...
- ML.DecTree.Forest: code for dealing with forests (collections) of decision trees **NOTE** This code should be obsolete now that ML.Composite.Composite is up and running.
- ML.DecTree.ID3: ID3 Decision Trees...
- ML.DecTree.PruneTree: Contains functionality for doing tree pruning
- ML.DecTree.QuantTree: Defines the class _QuantTreeNode_, used to represent decision trees with automatic...
- ML.DecTree.SigTree: Defines the class SigTreeNode, used to represent trees that use signatures (bit vectors) to represent data.
- ML.DecTree.Tree: Implements a class used to represent N-ary trees
- ML.DecTree.TreeUtils: Utilities for working with trees
- ML.DecTree.TreeVis: functionality for drawing trees on sping canvases
ML.Descriptors
- ML.Descriptors.CompoundDescriptors: descriptor calculator for compounds defined by a composition alone...
- ML.Descriptors.Descriptors: Various bits and pieces for calculating descriptors
- ML.Descriptors.MoleculeDescriptors: Various bits and pieces for calculating Molecular descriptors
- ML.Descriptors.Parser: The "parser" for compound descriptors.
ML.EnrichPlot: Command line tool to construct an enrichment plot from saved composite models Usage: EnrichPlot [optional args] -d dbname -t tablename <models> Required Arguments: -d "dbName": the name of the database for screening -t "tablename": provide the name of the table with the data to be screened <models>: file name(s) of pickled composite model(s).
ML.FeatureSelect
- ML.FeatureSelect.CMIM: for the moment this is using Francois Fleuret's cmim library to do the feature selection Reference: F.
- ML.FeatureSelect.rdFeatSelect: Module containing functions for feature selection
ML.GrowComposite: command line utility for growing composite models **Usage** _GrowComposite [optional args] filename_ **Command Line Arguments** - -n *count*: number of new models to build - -C *pickle file name*: name of file containing composite upon which to build.
ML.InfoTheory: Information Theory functionality
- ML.InfoTheory.BitClusterer
- ML.InfoTheory.BitRank: Functionality for ranking bits using info gains **Definitions used in this module** - *sequence*: an object capable of containing other objects which supports __getitem__() and __len__().
- ML.InfoTheory.entropy: Informational Entropy functions...
- ML.InfoTheory.rdInfoTheory: Module containing bunch of functions for information metrics and a ranker to rank bits
ML.KNN: Here is the implementation for K-nearest neighbors
- ML.KNN.CrossValidate: handles doing cross validation with k-nearest neighbors model...
- ML.KNN.DistFunctions
- ML.KNN.KNNClassificationModel: Define the class _KNNClassificationModel_, used to represent a k-nearest neighbhors classification model...
- ML.KNN.KNNModel: Define the class _KNNModel_, used to represent a k-nearest neighbhors model
- ML.KNN.KNNRegressionModel: Define the class _KNNRegressionModel_, used to represent a k-nearest neighbhors...
ML.MLUtils
- ML.MLUtils.VoteImg: functionality for generating an image showing the results of a composite model voting on a data set...
ML.MatOps: Matrix operations which may or may not come in handy some day...
ML.ModelPackage
- ML.ModelPackage.Packager
ML.NaiveBayes: An implementation of the Naive Bayes Classifier
- ML.NaiveBayes.ClassificationModel: Defines Naive Baysean classification model...
- ML.NaiveBayes.CrossValidate: handles doing cross validation with naive bayes models...
ML.Neural
- ML.Neural.ActFuncs: Activation functions for neural network nodes Activation functions should implement the following API: - _Eval(input)_: returns the value of the function at a given point - _Deriv(input)_: returns the derivative of the function at a given point The current Backprop implementation also requires: - _DerivFromVal(val)_: returns the derivative of the function when its value is val In all cases _input_ is a float as is the value returned.
- ML.Neural.CrossValidate: handles doing cross validation with neural nets This is, perhaps, a little misleading.
- ML.Neural.NetNode: Contains the class _NetNode_ which is used to represent nodes in neural nets **Network Architecture:** A tacit assumption in all of this stuff is that we're dealing with feedforward networks.
- ML.Neural.Network: Contains the class _Network_ which is used to represent neural nets **Network Architecture:** A tacit assumption in all of this stuff is that we're dealing with feedforward networks.
- ML.Neural.Trainers: Training algorithms for feed-forward neural nets Unless noted otherwise, algorithms and notation are taken from: "Artificial Neural Networks: Theory and Applications", Dan W.
ML.SLT
- ML.SLT.Risk: code for calculating empirical risk
ML.ScreenComposite: command line utility for screening composite models **Usage** _ScreenComposite [optional args] modelfile(s) datafile_ Unless indicated otherwise (via command line arguments), _modelfile_ is a file containing a pickled composite model and _filename_ is a QDAT file.