Machine Learning – Spencer Payne

Models built for client-style problems and scored on hidden holdout sets. Each write-up covers the approach, the honest results, and what I would change next.

The original Jupyter/Colab notebooks, including a from-scratch transformer language model in PyTorch and an RNN text-generation walkthrough, are in Machine_Learning/notebooks.

Who Will Subscribe? Bank Marketing Decision Tree

Machine Learning · Classification

A decision tree classifier predicting which customers will subscribe to a term deposit, used to target a bank’s phone campaign.

Forecasting Bike-Share Demand with a Neural Network

Machine Learning · Deep Learning

Python

TensorFlow/Keras

Neural Networks

Hyperparameter Tuning

A Keras neural network predicting hourly registered riders from weather and calendar data, tuned with grid search.

Predicting King County Home Prices

Machine Learning · Regression

Feature engineering and an XGBoost regressor that explains about 90% of the variance in home sale prices on held-out data.

Traffic Sign Recognition with a CNN

Machine Learning · Computer Vision

A convolutional neural network that classifies 43 types of German traffic signs, reaching about 88% validation accuracy with data augmentation.

Categories

Who Will Subscribe? Bank Marketing Decision Tree

Forecasting Bike-Share Demand with a Neural Network

Predicting King County Home Prices

Traffic Sign Recognition with a CNN