Python Kernel Quantile Regression
Explore the details of my project on Python kernel quantile regression. This project implements a kernel-based method for quantile regression in Python, providing a flexible and robust approach to modeling complex data relationships. This work resulted in a open-access and free of charge publication in the ACM SIGEnergy Energy Informatics Review.
For more information, visit the project repository on GitHub and see the paper:
Mail Marketing Database
Designed ER/EER, derived Relational Model schema and translated the schema into a MySQL database for a mail marketing company. Beside actual implementation a motivation is also provided explaining the specific design choices.
For more information, visit the project repository on GitHub:
Text and Spatial Analysis of Swiss Parliamentary Speeches
This report examines the Swiss parliamentary speeches across various dimensions. Through text and spatial analysis techniques, further insights on languages, cantons, congresspeople, factions and topics covered have been extracted. A variety of datasets have been employed throughout this project; the most important between all of these is the SwissParliamentarySpeeches dataset, this file stores all speeches given in the Swiss Parliament since 1999.
For more information, visit the project repository on GitHub:
https://github.com/luca-pernigo/swiss_parliament_speeches_analysis
English Premier League Dashboard
Developed an interactive dashboard for the english premier league. The dashboard drills down on various metrics and statistics providing interesting insights.
For more information, visit the project repository on GitHub:
Deep Learning
Four deep learning projects using the PyTorch framework.
First: Building a neural network learning a polynomial regression for the input data.
Second: Building a convolutional neural network to classify images from the CIFAR-10 dataset.
Third: Building a recurrent neural network for text generating given a input prompt.
Fourth: Building a transformer to solve simple mathematical questions.
For more information, visit the project repository on GitHub:
Monte Carlo Seminar Course
Source code, pdf report and slides for the Efficient Computational Algorithms seminar course. The topic I was assigned was Monte Carlo methods.
For more information, visit the project repository on GitHub:
Financial Modelling Challenge
The challenge was about creating a fictional fund and developing its strategy from scratch.
Our portfolio has been created and managed with Bloomberg. Moreover, we also used Excel for data analysis and to compute metrics. Finally, we used Python to automate repetitive stuff.
Beside designing and monitoring our portfolio we also edited reports for investors.
The strategies we implemented were:
- Discretionary Investing
- Quantitative Investing
- Macro Investing
For more information, visit our fund website:
Artificial Intelligence
Four artificial intelligence projects.
First: Implementing the Iterated Local Search algorithm for the Travelling Salesman Problem.
Second: Implementing Genetic Algorithm to optimize De Jong functions.
Third: Implementing Ant Colony System to solve the Travelling Salesman Problem.
Fourth:Implementing Monte Carlo Tree Search for the Orienteering Problem with Time Windows.
For more information, visit the project repository on GitHub:
Ordinary differential equations
This project explores various aspects of Runge-Kutta methods. The project is composed by three blocks.
First: Implementing collocation methods of arbitrary order, with a particular emphasis on Gauss and Radau methods.
Second: Incorporating time step adaptivity by employing the embedded methods strategy.
Third: Investigating various approaches for solving the Runge–Kutta nonlinear system. Factors considered were: Newton versus quasi-Newton algorithms, direct versus iterative solvers, Jacobian construction, matrix-free methods, and the adoption of preconditioners.
For more information, visit the project repository on GitHub: