possible Student projects

Association Studies

This involves determining whether there is any association between variables. This will be shaped by the following considerations
No of variables:  do we want to determine association between two variables, two groups of variables, or does any group contain a very large number of variables?
Type of Variables: What kind of variables are involved? Are they continuous or discrete? Are the variables collected over time?
Model Approach: What kind of statistical techniques should be used?
Interest: Do you have a data set of interest and want to find appropriate methods for it? Do you want to learn different methods and find data sets to practice the methods?

Image classification

Image classification has gained importance and popularity due to the rise of artificial intelligence along with other things.  This project involves studying different neural network architectures and using them for the classification purpose. Python is commonly used for these methods and prior knowledge would be useful.

text classification (natural language processing)

Natural language processing  is a field of designing methods that take unstructured natural language data as input or output. Unstructured text is everywhere, such as emails, chat conversations, websites, and social media but it’s hard to extract value from this data unless it’s organized in a certain way. Doing so used to be a difficult and expensive process. Text classifiers with NLP have proven to be a great alternative to structure textual data in a fast, cost-effective, and scalable way. 

High dimensional methods

Many studies involve a very large number of variables. Analyzing such data sets involve different statistical techniques. This project involves studying/investigating popular high dimensional methods.

deep learning for functional data

Functional data is exactly as it sounds – it consists of functions! It is basically collected over time/space. How can sequential networks such as RNN, LSTM  developed for time – series like data be used for functional data?  You will learn several neural network architectures along with their implementation. you will get exposure to a different data type.