Here You can find my projects which I prepared for my Machine Data Analysis labs.
-
Segmentation -> A program that will allow segmentation of car models, using clustering methods: KMeans, GMM, DBSCAN,
-
SVD -> Singular Value Decomposition represented at trivial example and at the image,
-
Inertion-Silhouette -> A program that thanks to inertion and silhouette score will help to decide what is the most optimal number of colors to reproduce the image,
-
PCA -> Principal component analysis, represented at wine data set, representation how PCA works.
-
TransferLearning -> The default selected network is Xception, train the network, train gradually thawing layers until you do not see overfitting. Check how the model works on the selected images. The code should display the image, tell you what its actual class is, and return the probabilities predicted by the model.
-
Autoencoder -> Add noise to any images loaded from disk. Build and train an autoencoder whose purpose is to remove noise from images. Visualize how the autoencoder works - show the noised and cleared image.