Copyright © 2022 Alessio Borgi
PROJECT SCOPE: Main PC Components Analysis through Statistics and ML Techniques.
PROJECT RESULTS:
- Data collection (CPU, Memory, Physical Disk, Cache, and Battery) through the Windows' "Performance Monitoring" application over multiple PCs.
- Clustering Techniques application (K-Means Clustering, K-Means++, GMM, K-Means Elkan Algorithm) classifying the data collected either idle or working.
- By-Hand K-Means and K-Means++ implementation.
- Elbow Method, Silhouette Coefficient, and AIC & BIC evaluation Criteria determining the best number of clusters.
- Best Algorithm choice based on Accuracy and Time Performances.
- Best PC choice based on Time Performances.
- Average Page Faults measurements (using Non-Parametric and Parametric Bootstrap).
- Anova-one-way-Test for Page and Cache Faults comparison test.
- Pairwise T-Test for finding similar Pcs’ behaviours.
- Linear Regression Application to determine "%Disk Time" vs "Average Disk Write Queue Length", and "Data Maps/Sec" vs "Copy Reads/Sec" relationships.
- Highly Potential Influence points individuation with Leverage and Cook's Distance Techniques.
PROJECT REPOSITORY: https://github.com/alessioborgi/Performance_Monitoring_Statistics_and_ML