23 projects, 2022 to 2026

Data work first. The five with plots are shown as figures; everything, coursework included, is in the index below them.

Daily TFT forecasts for the best, median and worst venues (MAE 9.23, 33.85 and 70.82). Crosses mark days outside the 80% interval.

Fast Food Demand Forecasting with Uncertainty Quantification

2026My MSc thesis: order forecasts for 17 Vilnius restaurants with ARIMA, DeepAR and a TFT, then a test of whether their prediction intervals hold.

Venues: 17
POS records: ~12.9M
Models compared: 4
TFT daily MAE: 37.3
Total80% interval coverage: 57%

A random walk with drift forecast of the gold price: the intervals widen the further ahead it looks.

Statistical Analysis & Forecasting

2024Descriptive and predictive analytics in R: PCA, clustering, ARIMA and ETS forecasting and multivariate modeling, all reported through RMarkdown.

Covers: PCA, k-means, ARIMA, ETS
Reports: RMarkdown

For this sample the model favors Neutral (0.68) over Emotional (0.28); the table lists the features behind that call.

Mathematical Methods of AI

2024A grab-bag of mathematically grounded AI work: explainable AI, recommender systems, Kalman filters and LDA topic modeling, with a focus on interpretability.

Covers: XAI, Kalman filter, LDA, recommender
Priority: Interpretability

Cotton crop separates cleanly on these two features; the gray soil classes overlap.

ML Lab Projects

2024Supervised learning labs across regression, classification (LDA, QDA, Logistic), KNN, and model evaluation.

Models: Linear, logistic, LDA, QDA, KNN
Evaluation: Cross-validation

Sunspot data as two lagged inputs against the output: most points sit near zero, a few cycles reach 150 and above.

AI Lab Methods & Projects

2024Lab-based AI projects covering ANN classification, Naive Bayes spam detection, time-series forecasting, and statistical data preprocessing.

Covers: ANN, Naive Bayes, time series, preprocessing

Index

Topic

Showing 23 of 23