PhD: Data Mining and Machine Learning

On July 15, 2018

University of Nottingham

The University of Nottingham is a public research university in Nottingham, United Kingdom. It was founded as University College Nottingham in 1881, and was granted a royal charter in 1948. The University of Nottingham belongs to the elite research intensive Russell Group association.

School of Computer Science

The School of Computer Science at the University of Nottingham is a modern and dynamic place to work and study. We providing state-of-the-art facilities and undertake world leading research. The Computer Science building is based in the centre of the picturesque Jubilee Campus. Just a short walk or 5 minute free bus from the renowned University Park Campus.

Subjects that I have Learned

While I was at the University of Nottingham, I took the opportunity to attend many lectures as they were free of charge of the PhD students while they were not compulsory.The subjects that I have attended are:

  • Academic English.
  • Academic Writing.
  • Advanced Mobile Programming.
  • Functional Programming with Python.
  • Human Computer Interaction.
  • Artificial Intelligence.
  • Introduction to Statistics.
  • Computer Graphics.

Algorithms I have Developed

In the process of my PhD, I have developed some algorithms that I have used in my research. The algorithms can be used in the data classification and data mining projects. Especially, for classifying behaviour over time and measuring changes in behaviour. The algorithms are:

Temporal RuleBased Classification

The proposed method for rule-based temporal classification has an advantage of classifying stocks according to a set of loosely defined rules presented by human experts without the need for a training data set. Temporal RuleBased Classification

Calculate Change Overtime

Calculate Change Overtime is to measure behaviour differences of items between time points requires three steps: The first step is to address time points, the second step is grouping similar behaviour and the last step is to find and measure the amount of differences between these time points. These steps will be addressed in the following paragraphs, and will be implemented and tested on the data in the next chapter.

Stock Price Download

Stock Price Download code is written in R language reads historical stock market prices from free sources namely google and yahoo. It fetches historical temporal data between two dates for selected vector of stock symbols (tickers) and return it as data.frame.

Stock Market Classification

Stock Market Classification uses the Temporal RuleBased Classification algorithm, this code classifies stocks according to a set of loosely defined rules presented by human experts without the need for a training data set. The classification is done on the basis of the stock price change over time. The data which has been classified is collected by using Temporal RuleBased Classification algorithm.

My Dissertation

You can find the latex version of my Dissertation in the GitHub repository, with a compiled PDF version. And the defense presentation is available here.


The certificate of my PhD is available here.


I have worked under the supervision of two of the most experienced researchers in the field of data mining and machine learning. I have also worked with a number of researchers in the field of artificial intelligence. My direct supervisor is Dr Uwe Aickelin and my co-supervisor is Dr Christian Wagner.


Polla Fattah