Team VBAP: A Kaggle Challenge

Project Description

This project is part of the DSA Major capstone which server to simulate a proper job experience. This included things like utilizing SCRUM and AGILE as well as working directly with a client. The goal we were assigned was to sign up for a competition on Kaggle and compete to reach at least the top half of the leaderboard. We ended up competiting on a competition entitled Hedge Fund - Time Series Forecasting which was meant to utilize Time Series Forecasting models and machine learning to predict data on a test set.

Meet the Team

Peter Davis: Project Manager and Data Modeler
Brandon Faulkenberry: Client Liaison and Data Analyzer
Austin Pham: Project Documenter and Data Visualizer
Victor Ojienabor: Project Documenter and Data Modeler

Peter Davis

Brandon Faulkenberry

Austin Pham

Victor Ojienabor

Client Information

Client Name: Dr. Cengiz Gunay
Industry: Professor at GGC
Goal of the project: Place within top 50% of the leaderboard

Dr. Cengiz Gunay

Dataset

Our Dataset Summary

Time Series Index Summary

Data Processing & Model

All of our work is stored within a google colab notebook

Solution Examples

LightGBM Model

LightGBM:This is an example of the decision tree created by our LightGBM model submission

LSTM Model

LSTM:This graph shows how the LSTM model predicted values compared to the actual values

GRU Model

LightGBM:This visualization shows how well the GRU model was able to predict values compared to the actual values

Our Placement!

Our Placement on Kaggle

Our placement:We placed 280 out of 983 which is in the top 30%. Well above our goal!