Predict.Transactions: Predict.Transactions

In mRcSchwering/abacus: abacus

Description

This function uses a machine learning method to classify each type of transaction for a Transactions object as created with Read.Transactions for example. A table with the actual transactions and the predicted types is returned.

Usage

## S3 method for class 'Transactions'
Predict(x)

Arguments

x	Transactions object (created with Read.Transactions for example)

Details

1. If there are new accounts (not existent in database yet), these accounts are written into the database. Also, if the user did make changes to these accounts, these changes are applied to the actual transactions accordingly.

2. A complete transactions table is created by combining the original transactions table with all information about the accounts. Numeric values are converted to Integers in this step (* 100).

3. If a Prediction Model was already created, it is read from the database

4. Feature Extraction and Prediction are done using the model

Value

Transactions object, a list of 4 elements:

• Transactions a data.frame of the file that was read

• Reference a data.frame of the reference account

• db character of database used

• Prediction a data.frame of transactions with additional columns of account information. It includes a column with the predicted type of each transaction.

See Also

Other procedures: Duplicated.Transactions, Duplicated, Enter.Transactions, Enter, Evaluate_Predictor, Predict, Read.Transactions, Read_csv, Read, Update_Predictor

Examples

db <- "test.db"
Create_testDB(db)

f <- system.file("extdata", "test_transactions.csv", package = "abacus")
cols <- list(name = 6, iban = 7, bic = 8, date = 3, 
             reference = 5, entry = 4, value = 9, currency = 10)
tas <- Read_csv("giro", f, cols, db)
tas <- Read(tas)

params <- list(
  nFeats = 200, 
  DDL = FALSE, 
  time = list(start = as.Date("2010-1-1"), end = as.Date("2011-1-1"))
)
InsertBLOB("Params", params, db)
Update_Predictor(db)

pred <- Predict(tas)

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