MachineLearningPractice

This project is all the different things I have been trying to get better at machine learning

Content List

• 📈 Linear Regression: I implemented my first model with Linear Regression to alter the gradient and train my model to produce the necessary outputs
• ⚪ Classification Problems: Working on solving a Neural Network Classification Problem instead of simple Linear Regression

Personal Project

• With the learning from the practice I did above, I created a model that was trained to determine whether a basketball player performed well or poor on a certain game night. I decided to train the model with Klay Thompson's 2022-2023 stats and it actually worked out!
• Here is some of the data I received so far

⚪ Table of Epoch Count vs Accuracy, Loss and Test Loss/Accuracy

Epoch: 000 | Loss: 0.69598, Accuracy: 50.00% | Test Loss: 0.81133, Test Accuracy: 68.00%
Epoch: 200 | Loss: 0.26894, Accuracy: 88.18% | Test Loss: 0.42585, Test Accuracy: 88.00%
Epoch: 400 | Loss: 0.15111, Accuracy: 94.55% | Test Loss: 0.54692, Test Accuracy: 92.00%
Epoch: 600 | Loss: 0.03830, Accuracy: 98.18% | Test Loss: 0.90846, Test Accuracy: 92.00%
Epoch: 800 | Loss: 0.00544, Accuracy: 100.00% | Test Loss: 1.62044, Test Accuracy: 88.00%

• NOTE: I think more data is needed to get a better test accuracy OR a complete rework to the neural model is needed

📈 Accuracy Graph

📈 Losses Graph