Data Preprocessing

 

***Note***

Due to the Coronavirus, many things are unclear. For the period that the university is closed because of Coronavirus, we will have online lectures.

Registration

The registration is carried out by the central registration process in January 2021.

You will be informed about the first meeting in the weeks after the registration closes.

Seminar Details

The learning objective of this pro-seminar is twofold, first to train students to improve their critical thinking through understanding research methodology and instructing them on how to write a scientific paper. Also in this pro-seminar, we will employ data preprocessing as a group of algorithms that enables students to become more familiar with data analysis and data science.

This pro-seminar briefly introduces fundamentals of data preprocessing which is one of the main steps in any data science project that involves transforming raw data into an understandable format. Real data is generally noisy, incomplete, inconsistent, and/or lacking certain behaviors. Data preprocessing comprises a series of techniques that resolve issues and their causes in order to produce more accurate and trustable results.

The following major topics are selected as working materials during this course:

Data cleaning/cleansing

• Noisy Data
• Missing Values

Data Integration

• Redundancy and correlation Analysis
• Data Value Conflict Detection and Resolution

Data Reduction

• Attribute Subset Selection
• Clustering

Data Transformation

• Normalization
• Discretization
• Decision Tree

Concept

Each participant will work on one of the above topics and summarize it in a short paper (maximum 15 pages), using LaTeX completed with an oral presentation of 20 minutes.

All topics will be introduced in the introductory session at the start of the pro-seminar. Participation is mandatory throughout the course. In the introductory session, topics will be assigned to the students and the deadline for submitting the presentations will be discussed.

The language of the course is English; therefore, both the presentations and summaries shall be prepared in English.

Prerequisites

Basic computer science skills and knowledge.

References

References for the specific topics will be given in the first session of pro-seminar.