Project case study · Machine learning development
IPL Score Predictor AI
A deep neural network built with Keras to estimate the final score of an ongoing IPL innings from live match context.
What it is
The IPL Score Predictor estimates an innings' final score from its current match state.
It turns changing match variables into an approachable prediction workflow backed by a Keras neural network.
Problem and approach
Problem
An innings evolves continuously, so a useful estimate must account for the score, resources remaining, recent performance, and the teams involved.
Solution
Represent the current match context as model inputs and use a trained deep neural network to estimate the likely final total.
How it works
Match context
The application collects runs, wickets, overs, recent performance, and team information.
Model inference
A Keras model processes the current context and predicts the final score.
How I built it
- 1.
Feature preparation
Selected and prepared match-state inputs that influence an innings' final score.
- 2.
Model development
Built and evaluated a deep neural network with Keras.
- 3.
Prediction interface
Connected model inputs and output to a practical score-estimation experience.
Challenges and decisions
Representing a changing match in a form a model can use consistently.
Approach: Combined current score information with recent-performance and team context.
Lesson: Feature design is as important as model architecture for applied prediction problems.
Skills and technologies
What I learned
- How to prepare structured inputs for a neural network.
- How to turn a trained model into an interactive application.