IS ZC464 - Machine Learning | Programming Assignment
Classification
© BITS-Pilani, 2017
This programming assignment is on classification. In particular, you have to implement one of the approaches for classification in any of the programming language of your choice.
Following are the sequence of activities involved.
(1) You are given a data set (a classic dataset, iris flower data set) where each observed example has a set of features and has labels. Labels are essential for learning any supervised learning algorithms. Details on the data set is available on https://en.wikipedia.org/wiki/Iris_flower_data_set.
The first task is to understand the data set and understand the classification problem posed on this data. You are free to store this data set in any format for your implementation, however you are not allowed to modify the data.
(2) Split the data set into training and testing sets. Every third row in the given data set is your testing set and the rest of data makes the training set.
(3) Choose one classifier (from discriminant based / instance based / likelihood based approaches) and implement this classifier, learn the necessary parameters using the training set.
(4) Classify the testing instances with the model built. Compute performance metrics and explain how well the model fit to the given classification problem.
What to Submit?
(1) Implementation Files, along with the input files in a folder named ‘implementation’
(2) A text file, detailing the software requirements to run your program, along with the instructions to run this.
(3) A word document, explaining your implementation, and details of activities from (1) to (4). Name this as report.pdf
(4) Zip items (1) to (3) in a file and name it as ‘your ID-Name.zip’ (All caps) and submit through the documentation
Other details
(1) However, there is no limit on the programming language that you use to implement the assignment, we recommend using C/Java/Python/MatLab.
(2) Recommending algorithms Decision Tree/ LDA/ Logistic Regression/ Naïve Bayes, however, you are free to make a choice.
(3) Ensure that you do not submit any downloaded codes. Write your own implementation. Submit the assignment even you have completed a part of it, and this will be evaluated.
Weightage is 10 % - Number of Days Given to Solve is 10 days – Deadline is 04-04-17
Classification
© BITS-Pilani, 2017
This programming assignment is on classification. In particular, you have to implement one of the approaches for classification in any of the programming language of your choice.
Following are the sequence of activities involved.
(1) You are given a data set (a classic dataset, iris flower data set) where each observed example has a set of features and has labels. Labels are essential for learning any supervised learning algorithms. Details on the data set is available on https://en.wikipedia.org/wiki/Iris_flower_data_set.
The first task is to understand the data set and understand the classification problem posed on this data. You are free to store this data set in any format for your implementation, however you are not allowed to modify the data.
(2) Split the data set into training and testing sets. Every third row in the given data set is your testing set and the rest of data makes the training set.
(3) Choose one classifier (from discriminant based / instance based / likelihood based approaches) and implement this classifier, learn the necessary parameters using the training set.
(4) Classify the testing instances with the model built. Compute performance metrics and explain how well the model fit to the given classification problem.
What to Submit?
(1) Implementation Files, along with the input files in a folder named ‘implementation’
(2) A text file, detailing the software requirements to run your program, along with the instructions to run this.
(3) A word document, explaining your implementation, and details of activities from (1) to (4). Name this as report.pdf
(4) Zip items (1) to (3) in a file and name it as ‘your ID-Name.zip’ (All caps) and submit through the documentation
Other details
(1) However, there is no limit on the programming language that you use to implement the assignment, we recommend using C/Java/Python/MatLab.
(2) Recommending algorithms Decision Tree/ LDA/ Logistic Regression/ Naïve Bayes, however, you are free to make a choice.
(3) Ensure that you do not submit any downloaded codes. Write your own implementation. Submit the assignment even you have completed a part of it, and this will be evaluated.
Weightage is 10 % - Number of Days Given to Solve is 10 days – Deadline is 04-04-17
No comments:
Post a Comment