Skip to content

DonaldChung-HK/INST_0060_datastore

Repository files navigation

INST0060 - Research Template 1 - Group 12

About

  • This project compare 4 models' performance on a bank consumers dataset.
  • The four models are:
  • Logistic Regression
  • Random Forest
  • KNN
  • Fisher Linear Discriminant

Installation

We provided a separate files with the required libraries to run the experiments: <requirement.txt> The steps to follow to create the relevant environment are in the Requirements section below.

Usage

To run the experiment use the following syntax on your machine's terminal: python model_comparison.py <file.csv> <target_value> <column_to_drop>

The experiment value options are:

  • Logistic_Regression ~ ( runtime 1 to 2 min)
  • KNN ~ ( runtime 1 min)
  • Fisher ~ ( runtime 1 to 2 min)
  • Random_Forest ~ ( runtime 1 min)

EXAMPLE: python model_comparison.py Churn_Modelling.csv Exited Random_Forest RowNumber,CustomerId,Surname

Content

The project is structure with:

  • model_comparison.py ~ main method which calls methods to fit and evaluate each models
  • model_fit folder ~ contains one file for each model. Each file has a main method called in model_comparison.py and the methods used to fit the models
  • fomlads.evaluate.eval_classification.py ~ contains methods to evaluate the models
  • fomlads.data.preprocessing.py ~ contains method used to pre-process the raw data file
  • fomlads.data.external.py ~ contains the methods used to standardise and normalise the data

Requirements

The requirements to run these experiments are contained in the <requirements.txt> file. This file should be used to create an environment following the steps below:

  1. conda create -n <name_of_your_environment> python=3.7
  2. conda activate <name_of_your_environment>
  3. python -m pip install -r requirements.txt
  • the requirements.txt file should be in the same directory your currently working on

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages