hraj10 Goto Github PK

bayesian-ml

For an assignment at LSE, I deployed and evaluated a number of Bayesian machine-learning techniques based on their capabilities to correctly distinguish benign from malignant tumours.

comparing-hyperparameter-optimisation-algorithms-for-ridge-and-lasso

For a project at the UvA, we compared different hyperparameter selection algorithms for the well-reknown regularisation techniques Ridge and LASSO regression.

crude-oil-forecasting

For my master's thesis, I investigate the predictive power of several linear and non-linear approaches for crude oil prices, among which linear time series models, neural networks, Gaussian process regression, and time-varying coefficient models.

modelling-health-expenditure---addressing-potential-endogeneity-with-the-cosslett-correction

For a project at the UvA, we aimed to estimate the impact of health insurance on health expenditures amongst men addressing potential endogeneity.

music-genre-classification-using-deep-learning

For a project at LSE, I compared multiple Deep Learning models to classify and eventually predict music genres based on 30-seconds audio segments.

rainforest-segmentation

A comparison of deep learning segmentation models for deforestation monitoring using multispectral satellite imagery. Research paper and project work completed as part of LSE's ST456 Deep Learning course.

temp-data