Posts

Motivation Why Not Regular Logistic Regression? Building Algorithm Motivation During my master studies, I was tasked to build a model that will predict whether a mobile ad will be clicked based on a large dataset from Kaggle. During the class, we learned Stochastic Gradient Descent (SGD) and Naive Bayes and hence those methods were supposed to be used in the assignment. We were also told that we’re going to struggle with the size of the dataset and it’s easiest if we implement the algorithms from scratch utilising the sparsity of the data....

Motivation Visualising Bitcoin Price Using Prophet Developing Backtesting Results Motivation Being passionate about finance and familiar with using Facebook Prophet for time-series forecasting, I decided it’d be a fascinating project to investigate the profitability of a bitcoin trading strategy relying solely on Facebook Prophet. There are many great reads about Facebook Prophet used to predict bitcoin price (e.g. here), but I haven’t found a post anywhere that would fit the model every day, make trades based on the output and backtest this strategy....

Motivation Simple data checks Statistical Analysis Conclusions What if the lottery was more complex? Motivation When exploring a few notebooks that analyze lotteries on Kaggle, I noticed that they mostly focus on interactions and correlations between numbers. However, I haven’t encountered a notebook that would analyze the occurrence of numbers using probability distributions. This notebook uses Python 3 to investigate this by comparing the occurrence of numbers with their theoretical distributions under the assumption that the lottery is fair....