Skip to content

josepmariall/x_biamow

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

55 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Predicting product attractiveness

This project aims at predicting the attractiveness of a household device.

Install

I provide the environment used to run this code.

Table of contents

The files contains the following main documents:

  1. Data Preparation (train)
  2. Data Visualizations and Data Wrangling
  3. Data Preparation (test)
  4. Data Preparation Train data (2)
  5. Algorithm selection
  6. Feature Selection, Algorithm fine tunning, Generalization evidence
  7. Running a prediction (inference

Some clarifications about the contents

File 2 clearly shows through certain visualizations the relationships between variables and the label, so we can see how 3 clearly distinct product types behave according to the variable.

File number 5 presents the results of running different algorithms.

File number 6 takes the best algorithm from file 5 and improves it, how? By selecting the best features selection for this model (hybrid method) and then tuning its parameters.

The author has done his best to produce an unbiased algorithm, however the algorithm is based on data, and data may be biased based on its collection process.

Code of ethics

This project has been undertaken complying with a code of ethics

License

This project is under Copyright © 2019 Josep Maria Niubo. It is free software, and may be redistributed under the terms specified in the LICENSE file.

Releases

No releases published

Packages

No packages published