sergiosjs / data_science_portfolio

A data science portfolio for academic and learning purpose.

Data Science Portfolio

Data Analysis and Visualisation

• Pokémon for Data Mining and Machine Learning (Kaggle):
  • Exploration
  • Storytelling
• Journeys made on service by the federal government of Brazil - 2019 (Dataset)
  • Exploration
  • [Storytelling]
• Exploiting Deep-Features Diversity in Food-11 Classification
  • Code
  • Report
• Survey: Open-set Classification
  • Survey file
• wiki4HE Data Set (Dataset)
  • Exploration
  • Storytelling

Docker Image to Data Science

Link

Other studies

• Maze Best Path: several state-space search algorithms applied to solve maze problems.
  • A*
  • Uniform Cost Search
  • Best First Search
  • Iterative Deepining Search
  • Depth Limited Search
  • Deep First Search
  • Breadth First Search
• Keras + MNIST: a simple test with Keras and MNIST to assess impact by varying hyperparans.
• Adaboost with numpy: Adaboost for categorical data implemented with numpy only.
• Reinforcement Learning: Some tests with reinforcement learning algorithms, including qlearn, sarsa, lambda sarsa, deep q learning
• Computational heuristics: Several algorithms of computational heuristics problems.
  • Hill Climbing
  • Iterated Local Search (ILS)
  • Greedy Randomized Adaptive Search Procedure (Grasp)
  • Tabu Search
  • Genetic Algorithm
• Running IRACE with python