1. Data Preparation

Workflow: Data Preparation

This workflow prepares the data for the next workflow ("My first Data Model") and uses some of the most common nodes for data preparation:

• Applying different strategies for missing values (Missing Value node)

• Creating subsets of the data (Row Sampler and Table Partitioner nodes)

• Shuffling (Shuffle node)

• Concatenation of data sets (Concatenate node)

• Normalizing data (Normalizer and Normalizer (Apply) nodes)

After preprocessing, the workflow writes the two subsets back to .csv files, one for the training set (top partitioning), one for test set (bottom partitioning).

Reading data

Missing value handling

• For all integer columns: replace with 0

• For "age": replace with mean value

• For "sex": remove row

• For "workclass": remove row

Creating subsets

Shuffling

Concatenate

Writing data

Normalizing data