Neural Network for House Price Prediction

Oct 2024 - Nov 2024

GitHub Repo

Synopsis¶

For my Year 3 ML module at Imperial College London I was tasked with:

• Implementing a decision tree algorithm, using it to determine an indoor location based on WIFI signal strengths collected from a mobile phone.

• Developing and optimising a neural network architecture to predict the price of houses in California using the California House Prices Dataset.

Skills¶

Below are the skills and experiences gained from the project:

• Implemented the following:

  • K-NN Classifiers
  • Decision Trees
  • Evaluation of ML models (Confusion matrix, performance metrics, etc)
  • Hyperparameter tuning (dataset split + crossvalidation)
  • Principles of overfitting and mitigation
  • NNs
    • Linear regression
    • Activation functions
    • Backpropagation
    • Gradient Descent
    • Overfitting

• Competency with the following libraries:

  • PyTorch
  • Numpy
  • Pandas
  • Scikit-learn