/loan-prediction-ml

3 layered neural network to predict the approval of a loan

Primary LanguageJupyter Notebook

Loan Prediction Neural Network

A simple 3 layered neural network trained on a Kaggle dataset by Bhavik Jikadara.

  • This is a simple solution averaging at around 67.74% accuracy.
  • The original dataset was modified since I do not understand how to work with labels (as of the day I'm writing this)
  • Rows containing null values were deleted in the data cleanup (again because I wanted to simplify this process for me)

This is one of my first ML projects and therefore might be extremely basic and innacurate. Be kind :)

Instructions:

Local Runtime using Python3

  1. Clone/download the files.
  2. Make sure you have Python3 and the following dependencies installed: tensorflow, pandas, numpy,, matplotlib, sklearn
  3. Run main.py using CLI command python3 main.py
  4. The runtime shouldn't take more than 20s (at most).

Jupyter Notebook (or) Google Colab

  1. Initialise Jupyter Notebook (or) Google Colab (for Colab, you need to connect to a runtime server)
  2. Make sure that your runtime server has Python3 and the following dependencies installed: tensorflow, pandas, numpy,, matplotlib, sklearn
  3. Upload all the files provided (you may skip the main.py file)
  4. Make sure to set the correct path of the loan_data.csv in loan_predictor.ipynb at dataset = pd.read_csv(<correct_filepath>)
  5. Run the loan_predictor.ipynb file, cell-by-cell.

Have a great day:)

- Priyanshu