A small-scale flask server facial recognition system, using a pre-trained facenet model with real-time web camera face recognition functionality, and a pre-trained Multi-Task Cascading Convolutional Neural Network (MTCNN) for face detection and cropping.
This project consists of implementations of several kNN algorithms for road networks (aka finding nearest points of interest) and the experimental framework to compare them from a research paper published in PVLDB 2016. You can use it to add new methods and/or queries or reproduce our experimental results.
TextureBasedImageRetriever a Content Based Image Retriever that focuses on texture. It implements the offline phase which is the calulation of descriptors of all images in the datasetn, and the online phase that return the n-similar images from dataset given an input image.
A Java console application that implements the factionality of the knn algorithm to find the similarity between a new user with only a few non zero ratings of some locations, find the k nearest neighbors through similarity score and then predict the ratings of the new user for the non rated locations.
Clustered customers into distinct groups based on similarity among demographical and geographical parameters. Applied PCA to dispose insignificant and multi correlated variances. Defined optimal number of clusters for K-Means algorithm. Used Euclidian distance as a measure between centroids.
A Java console application that implemetns k-fold-cross-validation system to check the accuracy of predicted ratings compared to the actual ratings and RMSE to calculate the ideal k for our dataset.
A Java console application that implements the factionality of the knn algorithm to find the similarity between a new user's location preferences and the locations. The binary data (0,1) are the location characteristics.
Sviluppo dell'algoritmo esteso di euclide. Permette all'utente di calcolare l'MCD tra due numeri interi e restituisce i coefficenti dell'identità di Bezout.
This course teaches you how to calculate distance metrics, form and identify clusters in a dataset, implement k-means clustering from scratch and analyze clustering performance by calculating the silhouette score
A K Nearest Neighbors classifier developed from scratch for self-learning purposes. Accuracy is off the charts, since we have full control on the algorithm.
