# ME781 - Engineering Data Mining and Applications

#### Instructor

Prof. Vinay Kulkarni and Prof. Asim Tewari

#### Course Difficulty

Moderate (An ideal introduction to the buzzword - Machine Learning! :D)

#### Time Commitment Required

Attending lectures would be enough to cover the theoretical base while actual understanding would only be acquired after solving the coding assignments. These weekly coding assignments would require approximately 3-5 hours per week depending on the topic and adeptness in Python coding.

#### Grading

Moderately strict (8 AAs and 7 ABs and in a class of approximately 73)

#### Attendence Policy

Attendance above 90% ensure full 5 points in the final evaluation, and a linear mapping was followed for attendance below 90%

#### Pre-requisites

An introductory course in coding (like CS101) is crucial while knowledge of Python is a plus although you can learn it as the course proceeds

#### Evaluation Scheme and Weightage

The evaluation was based on one quiz (which accounted for 10%), coding assignments (which accounted for 10%), mid semester examination (which accounted for 25%), end semester examination (which accounted for 20%), course project (which accounted for 30%) and attendance (which accounted for 5%). Good performance in the course project, therefore, was crucially important for a good grade.

#### Topics Covered in the Course

Essentials of Statistics, Introduction to Data Mining, Regression, Classification and Clustering (k-means and hierarchical) algorithms: Linear regression, Multiple linear regression, Logistic regression, Decision Trees, Random Forest, Bagging, Artificial Neural Networks, Support Vector Machines, Bayesian LDA & QDA, Concept of overfitting, Bias-variance tradeoff, Resampling and Cross-validation, Regularization, Principal component analysis, knn regression, Big Data Analytics, Techniques for storing and processing Big Data, Introduction to Keras - Convolution neural networks, ResNets, LSTMs

#### Mechanism of Instruction and Teaching Style

The instructor explains concepts well and clears all the doubts. He teaches with quite a good pace. Theory, in general, may seem difficult (especially the mathematical formulations of the algorithms but they aren’t very important :P) to comprehend but it all boils down to solving the assignments.

#### Assignments and projects in the Course

Understanding of concepts is almost entirely dependent on the sincerity with which you would attempt and complete the weekly coding assignments since the concepts require practical implementation rather than theoretical understanding. The course project is a very important aspect of the course. All the learnings culminate to the implementation of all the concepts that you’ve learned throughout the course for a particular case/situation (using a particular data set having practical applications). The code, project report and a viva-voce session encompass the grading scheme for the project. Nevertheless, the instructor’s objective is to only help you learn substantially well and therefore points out areas of improvement in the project during the viva-voce session and expects re-submission of the project with the due changes made. Such a practice is rare but should only be taken in a positive attitude!

#### Exams

The quiz, mid semester and the end semester examinations are easy if the course material is reviewed sincerely and the concepts (especially bias-variance tradeoff) are well understood. Expect the questions to be majorly numerical-based with heavy calculations (mostly calculating the statistical quantities). Understanding the solved problems in the slides is extremely crucial!

#### References

Although slides are enough, ‘An Introduction to Statistical Learning’ by James, Witten, Hastie, and Tibshirani can be helpful.

#### Importance of Course

Since Machine Learning is the name of the game in all aspects, be it business analytics or scientific research, this course can give you the right kick in that direction although there is a lot to learn in this field! Also, if you’re afraid of the mathematics heavy ML courses offered by the EE and the CSE department, then the ME department is here for your rescue! :P

#### Motivation to take the course

To learn ‘Machine Learning’ (due to a little peer pressure as well XD)

*Review by:* Dhvaneel Visaria