Libraries

Using the most popular libraries in data science:

• [scikit-learn] - Machine Learning Algorithms and helper methods.
• [numpy] - Simplifies linear algebra operations using C code for increased efficiency.
• [pandas] - Library for structured data manipulation.
• [matplotlib] - Visualization functionalities for data analysis.
• [nltk] - Natural Language Processing tool.
• [gym] - Reinforcement Learning applications.

Projects

Unsupervised Learning (K-means):

• Implementation of the k-means algorithm and evaluation using the elbow method, the silhouette coefficient and PCA.

Reinforcement Learning:

• Train an agent from the mini-grid game through reinforcement learning

Supervised Learning (Decision Trees):

• Implementation of a decision tree for continuous attributes and application to a real problem with a heavily unbalanced dataset.

Supervised Learning (Naive Bayes for text):

• Implementation of naive bayes to identify spam emails using stemming, tokenization and stopword elimination techniques.