Mark Borg

Blog


t-SNE and Machine Learning

t-SNE and Machine Learning

  This blog continues on my previous entry on using t-SNE for exploratory data analysis. Now we will consider t-SNE for use within a machine learning system. [read more]

Multi-Dimensional Reduction and Visualisation with t-SNE

Multi-Dimensional Reduction and Visualisation with t-SNE

  t-SNE is a very powerful technique that can be used for visualising (looking for patterns) in multi-dimensional data. Great things have been said about this technique. In this blog post I did a few experiments with t-SNE in R to learn about this technique and its uses. [read more]

Visualising PV output with 3D Surface Plots

Visualising PV output with 3D Surface Plots

  In order to experiment with the 3D plotting functionality of R, I used nearly a year’s worth of power output data from my solar panel installation at home. [read more]

Solving River-Crossing Puzzles with R

Solving River-Crossing Puzzles with R

  River-Crossing Puzzles are a popular class of puzzles in the field of AI. Many flavours of these puzzles exist. Here we use R to provide a somewhat generic framework to model and solve these type of puzzles. [read more]

Introductory Video on Search Spaces

Introductory Video on Search Spaces

  An important concept in the field of Artificial Intelligence is that of Search Spaces or State Space Search. [read more]

Big Data Summit Malta

Big Data Summit Malta

  I have attended the Big Data Summit, Malta, held on the 22nd of June 2016. There were a number of interesting talks at this first summit on this subject. It was worth its value. There was also time for networking and had some good discussions with both old colleagues and new contacts.   [read more]

Experiments in Optdigits Classification

Experiments in Optdigits Classification

  Optdigits is a well-known dataset consisting of a collection of hand-written digits available at the UCI Machine Learning Repository. In this blog I report classification results obtained using various machine learning classification techniques. [read more]

Kaggle Shelter Animal Outcome Competition

Kaggle Shelter Animal Outcome Competition

  In this blog I’ll go over some of my experiments on predicting the outcome of shelter animals. This work was done as my contributions to a competition on Kaggle. One particular focus of this blog entry is on data exploration and feature engineering. This dataset had some interesting variables that needed some work before they could be used as predictor variables. [read more]

Barcode Recognition

Barcode Recognition

  I worked on the BarcodeReader program as part of a set of Machine Vision demo programs that I did for some potential leads in the manufacturing industry. [read more]

Predicting Customer Satisfaction

Predicting Customer Satisfaction

  The first Kaggle competition that I participated in dealt with predicting customer satisfaction for the clients of Santander bank. My submission based on xgboost was ranked in the top 24% of all submissions. Kaggle ranks submissions based on unseen data – a public leaderboard dataset (test dataset with unseen targets) for use during development and a final private leaderboard dataset (completely unseen data) that is used for the final ranking of the competition. In this blog I will cover some aspects of the work done, in particular, initial data exploration and balancing the dataset to aid machine learning predictions (happy customers far outweighed unhappy ones). [read more]