This is the summary documentation of my course study of Machine Learning at Parsons' Data Visualization MS porgram in Spring 2022.
In the first five weeks, we practiced in a modular format and then we moved onto three machine learning projects.
- Week 01: Basic linear model fitting
- Week 02: Python practice
- Week 03: Basic machine learning fitting
- Week 04: Multiple models
- Week 05: Bag of Words
-
Project 1: Movie Reviews | Natural Language Processing
In this project, I experimented with multiple natural language processing tools to create a binary predictor of movie reviews. I worked with bag of words, Tfidf, sparse matrix, and text pre-processing techniques such as part of speech tagging and lemmatization. I used Ridged Regression and simple Bag of Words along with regulation to acheive the best outcome. Project Folder
-
Project 2: Image Recognition | Deep Learning
In this project, I experimented with various Skimage tools to process images and pass the image features into both Perceptron and multi-layer Perceptron Models. I acheived best model outcome by using Histogram of Oriented Gradients (HOG) and simple Perceptron. Project Folder
-
Project 3: Food Description Clustering | Unsupervised Learning
Drawing from lessons of the previous two projects, it seems that the simpler the model is the powerful the result is. In this project, I used a simple Bag of Words to extract food description features and KMean clustering. Project Folder