For Assignment 5, you will be developing classification models to predict which passengers survived the sinking of the Titanic. In the data section of this page, you will find a labeled training set (Lab5_train.csv) and an unlabelled test set (Lab5_test.csv). To start off the assignment, you will form groups of 2 to 5. As a group, you will then be selecting models (i.e. choosing an algorithm, tuning parameters, etc) using the training set, then submitting predictions for the test set in the specified format.
\r\nThe competition will be conducted in two phases. During Phase 1 (02/23 - 03/02), each group is required to make at least 4 submissions from at least 2 different algorithms (like svm, neural networks, decision trees, or others that we have not directly used in class - but no ensemble models). Each group is allowed two submissions per day.
\r\nDuring Phase 2 (03/03 - 03/12), each team should work on developing an ensemble model. You can start Phase 2 as soon as your team is happy with the individual algorithms, but make sure each group has completed the Phase 1 requirements first.
\r\nSee the evaluation section of this site for details on scoring.
\r\nThis is based on the Kaggle's popular Titanic competition.
", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Evaluation", "content": "The evaluation metric for this competition is accuracy (% correct). The data has been split 70/30 into a labeled training set and an unlabeled testing set. You will be submitting your predictions on the testing set. When you submit a set of predictions, you will receive an accuracy score based on a randomly-sampled 50% of the test set (the random sample will be constant across submissions). This score will show up on the competition's leaderboard. Your final score will be based on your accuracy on the whole test set, and will be released once the competition has closed.
\r\nEach group is limited to 2 submissions per day.
\r\nA sample submission csv file has been provided. Your submission should have headers (PassengerID,Survived) with two columns of output- the passenger id and the prediction for survival (1 for survived, 0 for did not survive). Kaggle will interpret each class prediction as a string ('1' vs '0'), so make sure your survived column does not include decimal places.
", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "rules", "content": "\r\nYou cannot sign up to Kaggle from multiple accounts and therefore you cannot submit from multiple accounts.
\r\nPrivately sharing code or data outside of teams is not permitted. It's okay to share code if made available to all participants on the forums.
\r\nTeam mergers are allowed and can be performed by the team leader. In order to merge, the combined team must have a total submission count less than or equal to the maximum allowed as of the merge date. The maximum allowed is the number of submissions per day multiplied by the number of days the competition has been running.
\r\nThe maximum size of a team is 5 participants.
\r\nEach group may submit a maximum of 2 entries per day.
\r\nStart Date: 2/23/2017 2:30 PM UTC
\r\nEnd Date: 3/12/2017 11:59 PM UTC
\r\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false } ] }