A collection of small, practical machine learning projects — each one built to solve a real problem rather than as a toy exercise.
A K-Nearest Neighbors model trained on Serbian optical internet plans (download/upload speed, price, bundle type, discount status) to recommend a good plan given a budget and speed requirement. This is the model that actually helped my dad pick a good internet provider.
A linear regression model that predicts mobile device usage trends over time, filtered specifically to mobile phones. The dataset it uses, mobile_devices.csv, is based on data from the Republički zavod za statistiku Republike Srbije (Statistical Office of the Republic of Serbia).
A decision tree classifier that predicts whether a bank customer will accept/subscribe to a credit offer, based on features like age, job, education, loan status, and previous campaign outcome.
A linear regression model tracking SSD vs. HDD price trends over time to project future pricing.
A K-Nearest Neighbors–based recommendation model for SSDs, matching drives by capacity, interface, form factor, drive class, media type, and production status.
- The
ML_projects(No datasets)folder contains the notebooks without their underlying data files. - Most notebooks use
scikit-learn(KNN, linear regression, decision trees) withpandasfor data handling andpicklefor model persistence.
See LICENSE.