Master of Science in Computer Science

Faculty / School

School of Mathematics and Computer Science (SMCS)


Department of Computer Science

Date of Submission



Dr. Zaheeruddin Asif, Assistant Professor, Department of Computer Science, Institute of Business Administration (IBA), Karachi

Project Type

MSCS Survey Report


Regulatory Compliance is organization’s conformation to some laws and regulations relevant to its business. These laws are created by regulatory authorities like State Bank of Pakistan creates regulatory laws for financial authorities in Pakistan. Currently, all financial institutions create compliance programs to handle the compliance of the organizations where manual work is done for identifying non-compliance and controlling it.

To automate this, we can use the Natural Language processing which is a branch of Artificial Intelligence through which computer can understand human language and speech. RegTech is the application of emerging technologies in field of regulatory processes to make it more enhanced and automated.

This research focuses on one of the applications of Natural Language Processing in the field of RegTech i.e. information extraction techniques to enhance regulatory compliance. Information Extraction is the branch of natural language processing used to find relevant information in unstructured text. To find the relevant part that what one organization need to comply can be achieved by using information extraction techniques i.e. Sentiment Analysis, text summarization etc.


As different information extraction techniques were found in this research it can be said that information extraction can surely enhance the way of regulatory compliance checking and is also enhancing currently but in a limited way. Talking about which technique is best in all of these then each and every technique has its own benefits depending upon the situation but ontology based semantic approach is most widely used technique for the information extraction specially when domain is specific. Machine learning approach can be used as a part of the whole model to classify concepts and auto populate ontology.

The full text of this document is only accessible to authorized users.