This repository contains project for the Time Series Analysis course conducted at Big Data Analysis at WUST.
Based on Human Activity Recognition Data Set (source: https://archive.ics.uci.edu/ml/datasets/human+activity+recognition+using+smartphones ), which basically are 9 types of signals collected on 6 recorded activities, the goal is to transform them using Continuous Wavelet Transform and create machine learning model (CNN created using PyTorch) being able to classify position (activity) on transformed dataset.
As an additional experiment, dataset was passed to multi-input ResNet50 model to compare quick results based on single training.