Machine Learning

This was the final group project graded for Algorithmic Machine Learning, one of the 3 graduate courses during my last semester of Bachelor's and part of my Accelerated Master's in Information Technology and Analytics (MITA).   

My group of 3, including me, applied several machine learning methods to build a pixel-level depth regression model using an RGB-D dataset. Specifically, we implemented and compared the following approaches:

• Linear and Polynomial Regression – to establish baseline performance and capture basic depth–color relationships.

• Random Forest Regression – to model nonlinear dependencies using an ensemble of decision trees.

• Support Vector Regression (SVR) – to handle complex patterns in the RGB-to-depth mapping through kernel-based learning.

• Neural Network Regression – to learn deeper feature representations for improved depth prediction accuracy.

• K-Nearest Neighbors (KNN) (not discussed in class) - to estimate depth by averaging values from the closest neighboring data points in the RGB feature space, leveraging local similarity for non-parametric regression.

We applied these techniques, according to the concepts and methodologies discussed in class, adapting each model appropriately.

**The code file can be seen in Google Colab. **

This was a personal project to learn machine learning in Python. I chose, what I felt, was the most easiest dataset to work with - stock data. 

I then used a classification prediction model (specifically random forest) to predict future stock prices. 

The predictions are inaccurate, and I would not base stock trading on the visualizations. So far, it seems to simply repeat the superficial patterns found in the historical data. 

The point was simply to get experience with Python sklearn libraries.

Link to project is within the captions to the right. -->

On 7/20/2025 (Sunday), predict prices until 7/23/2025 (Wednesday), based on 7 days of stock data with 5-minute intervals

This was a graded project for Data Mining for Business Intelligence. 

I cleaned, manipulated, and analyzed (through data visualizations and models) a wine quality dataset from Kaggle with R. 

The project required using at least 7 models for classifying, clustering, and association analysis.  

Finally, we had to record a presentation and submit. 

As part of the (Professional) Data Scientist with R certification on DataCamp, I had to submit a practical with a presentation. I can't show the presentation recording, but I can show the slides along with the written report if that's okay with you.  

To the left are the instructions for the practical. Below are the written report and the presentation used for the submission. Unfortunately, I can't show the presentation recording used, though. The point of the project was to classify each recipe as one that brough high-traffic to company website or not.  

**I cannot share links with those who do not have a DataLab account, so I recorded it in a video instead. 

As part of the Data Scientist (Associate) with R certification on DataCamp, I was required to complete the code for an unfinished written report on predicting house sales based on given inputs. 

**I cannot share links with those who do not have a DataLab account, so I recorded it in a video instead.