AI Productivity Predictor
Features (raw model) Ai classifcation models, from scratch to predict worker productivity based on worker inputs, using a custom made neural network taht utillises sigmoid function and back propogation.
Transforms raw productivity data into a binary classification problem
Handcrafted Neural Network. Built from scratch using NumPy—no frameworks
Dynamic backpropagation with customizable learning rate
Tracks F1-score, precision, recall, and accuracy to gauge model effectiveness
Plots a loss-over-epochs graph
Features (Scikit Learn Model) Ai classifcation models, both from scratch and from SciKit learn to predict worker productivity based on worker inputs, using neural networks.
Uses scikit-learn for a polished, production-ready workflow (train-test splits, scaling, metrics).
Trains a neural network iteratively (warm_start=True) for fine-grained control over epochs.
Tracks loss and metrics (F1, precision, recall, ROC AUC) in real time.
Normalizes features to standardized scales for optimal neural net performance
Plots error reduction over epochs to visualize learning dynamics
Components
SciKit Learn
Sigmoid Function
Python
Back Propogation
Industry-Standard Model Analysis Metrics
Repo