BestVectorTable

A data.table with the missing values you wish to fill

The collum with the missing values you wish to fill

goal

What will be the length of possible combinations you will test example if 2 they will test up to all possible pairs of collums

maxi

Measure of error, the bigger the less likely you will get the right prediction

repetitions

When you pair all possible combination of tuples a percentage of them will show only once, trigger rejects the set if this percentage is higher than this value

trigger

Rejects collumns that the ratio of unique values to total values is higher than this value, primary keys have ratio equal to 1

ratio

<code>BestVectorTable</code> Asks for a data.table and some parameters and returns the best combination of collums to predict the missing value

How to deal with missing data?Based on the concept of almost functional dependencies, a method is proposed to fill missing data, as well as help you see what data is missing.
The user can specify a measure of error and how many combinations he wish to test the dependencies against, the closer to the length of the dataset, the more precise.
But the higher the number, the more time it will take for the process to finish.
If the program cannot predict with the accuracy determined by the user it shall not fill the data, the user then can choose to increase the error or deal with the data another way.

Rafael Pereira

cleanerR

How to Handle your Missing Data

BestVectorTable function

<code>BestVectorTable</code> Asks for a data.table and some parameters and returns the best combination of collums to predict the missing value — BestVectorTable

BestVectorTable: `BestVectorTable` Asks for a data.table and some parameters and returns the best combination of collums to predict the missing value

Description

Usage

Arguments