Package de.jungblut.reader

Class CsvDatasetReader

java.lang.Object

de.jungblut.reader.CsvDatasetReader

public final class CsvDatasetReader
extends java.lang.Object

Binary dataset reader from CSVs.

Author:

thomas.jungblut

Method Summary

Modifier and Type	Method	Description
static Dataset	readCsv(java.lang.String path, char separator, java.lang.Character quote, int outcomeIndex, boolean skipHeader, boolean cacheOutcomeVectors)	Reads a csv into feature and outcome arrays.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

readCsv

public static Dataset readCsv(java.lang.String path,
                              char separator,
                              java.lang.Character quote,
                              int outcomeIndex,
                              boolean skipHeader,
                              boolean cacheOutcomeVectors)

Reads a csv into feature and outcome arrays.

Parameters:

path - the path to read from

separator - the separator to use

quote - the quote, null if none present

outcomeIndex - the index of the outcome, everything else is considered a feature.

skipHeader - if true it will skip parsing the first line.

cacheOutcomeVectors - if true it caches the vectors based on the outcome value. This saves a ton of memory for classification problems that share only a couple of unique values.

Returns:

a new dataset.