FeatureExtraction: FeatureExtraction

In mRcSchwering/abacus: abacus

Description

This function does a feature extraction to create an analytics base table from the transactions and the personalAccounts tables. The features are mostly binary (occurence of a word) or discrete (count of a word) and very sparse.

Usage

FeatureExtraction(ta, pa)

Arguments

ta	data.frame of transactions joined with accounts as returned by Select("transactions", ...)
pa	data.frame of personalAccounts joined with accounts as returned by Select("personalAccounts", ...)

Details

The analytics base table (ABT) containes feature values as int. Columns represent features, rows data points (= transactions). The ABT is accompanied by a FeatureList which describes the features. It is necessary for conversion between different ABTs.

Value

list of 2 objects

• ABT int matrix analytics base table containing features as columns, data points as rows

• FeatureList data.frame description of ABT columns with names and values

See Also

Other machine learning: CV, Convert, Prediction, Training

Examples

db <- "db/test.db"
Create_testDB(db)
ta <- Select("transactions", db)
pa <- Select("personalAccounts", db)
res <- FeatureExtraction(ta, pa)

mRcSchwering/abacus documentation built on May 21, 2019, 9:18 a.m.