In this repository you may find data and code used for a machine learning project in sensor data done in collaboration with my colleagues Lorenzo Ferri and Roberta Pappolla at the University of Pisa.
This is an analysis of the 'Room Occupancy Dataset' with the purpose of predicting whether a person is present in the room or not solely based on the sensor data on the quantity of light, CO2, humidity, etc. Project was a compulsory part of the Data Mining II: Advanced Topics and Applications course at the University of Pisa.
In folders and subfolders of this directory, you may find Jupyter Notebooks dealing with this dataset in following respects:
-Advanced Classification (SVM,SVC, Convolutional and Recurrent NN, Deep NN,
Ensemble Classifiers…) & Clustering (K-Means, DB Scan, Transactional C.)
-Time Series Analysis & Forecasting (Dynamic Time Warping, Motifs, Shapelets, TS Classification & Clustering)
-Dimensionality Reduction with PCA, SVD, UFS, RFE…
-Sequential Pattern Mining (GSP Algorithm)
-Outlier&Anomaly Detection (ABOD, LOF, KNN, COF, INFLO, Grid-Based…)
-Principal Component Analysis (PCA)
-Explainable Machine Learning
-Imbalanecd Learning
Find the written report in the .pdf file in this repository.
Any contribution is welcome!