Machine Learning portfolio focused on practical experimentation with supervised learning, unsupervised learning, forecasting and model comparison using Python and real-world datasets.
This repository documents my hands-on learning journey through:
- predictive modeling
- classification
- clustering
- association rule learning
- time series forecasting
- machine learning forecasting
- feature engineering
- exploratory data analysis
- model evaluation and visualization
The objective is not only to train models, but also to understand:
- how models behave
- how they generalize
- how different approaches compare
- how Machine Learning can be applied to real-world problems
Comparison between supervised classification and unsupervised clustering using the Iris dataset.
The project explores whether clustering algorithms such as KMeans can identify natural flower groups without using species labels.
Association Rule Learning project using the Apriori algorithm to discover purchasing patterns and product relationships in supermarket transactions.
Classical forecasting project using the Air Passengers dataset.
Main concepts explored:
- trend
- seasonality
- stationarity
- ACF / PACF
- ARIMA forecasting
- residual analysis
Forecasting project transforming time series data into a supervised learning problem using lag features.
Models explored:
- Ridge Regression
- Random Forest
- XGBoost
Main concepts explored:
- lag engineering
- exogenous variables
- forecasting evaluation
- model comparison
- Python
- Pandas
- NumPy
- Scikit-Learn
- Statsmodels
- XGBoost
- Mlxtend
- Matplotlib
- Seaborn
- Plotly
- Jupyter Notebook
- Google Colab
Machine-Learning-Projects/
├── association_rules/
├── binary_classification/
├── multiclass_classification/
├── regression/
├── time_series/
├── unsupervised/
├── docs/
├── requirements.txt
├── README.md
└── LICENSE
| Section | Main Topics |
|---|---|
| Regression | Linear Regression, Random Forest, XGBoost |
| Binary Classification | Heart Disease, Mushroom Classification |
| Multiclass Classification | Iris, IoT Agriculture, Abalones |
| Unsupervised Learning | KMeans, DBSCAN, Hierarchical Clustering |
| Association Rules | Apriori, Market Basket Analysis |
| Time Series | ARIMA, Lag Features, ML Forecasting |
Each section includes:
- datasets
- notebooks
- visualizations
- model comparisons
- forecasting or evaluation workflows
Current areas of exploration include:
- advanced forecasting workflows
- SARIMAX
- XGBoost forecasting
- feature engineering
- model comparison
- forecasting evaluation
- machine learning experimentation
Bea Lamiquiz
Machine Learning portfolio focused on forecasting, model comparison, visualization and practical experimentation using real-world datasets.