scale_encode_combination's Introduction

scale_encode_combination

introduction

It helps with scaling & encoding of dataset.
This function creates 10 types of datasets that combine the scalers and encoders below.

Scaler	Encoder
standard	ordinal
minmax	onehot
robust
maxabs
normalize

This project belongs to Data science term project, Gachon Univ. ( 2021 spring semester )

Getting Started

Just clone or download!
Run the function by referring to the test.

example

import pandas as pd
from Scale_Encode_Combination import scale_encode_combination
from pprint import pprint

# sample dataset (6 row)
dataset = pd.read_csv("sample_datas.csv")

#set numerical feature list
numerical_feature_list = ['duration_ms']

#set categorical feature list
categorical_feature_list = ['time_signature']

# run!
combination_dataset = scale_encode_combination(dataset, numerical_feature_list, categorical_feature_list)

#check result
pprint(combination_dataset)

# [output]
# {'minmax_onehot':    Unnamed: 0  duration_ms  time_signature_1  time_signature_3  time_signature_4  time_signature_5
# 0           1     0.000000                 1                 0                 0                 0
# 1           2     1.000000                 0                 0                 0                 1
# 2           3     0.943277                 0                 1                 0                 0
# 3           4     0.777565                 0                 0                 1                 0
# 4           5     0.967558                 0                 0                 1                 0
# 5           6     0.434648                 0                 0                 0                 1,
#  'minmax_ordinal':    Unnamed: 0  duration_ms  time_signature
# 0           1     0.000000             0.0
# 1           2     1.000000             3.0
# 2           3     0.943277             1.0
# 3           4     0.777565             2.0
# 4           5     0.967558             2.0
# 5           6     0.434648             3.0,
# ....