Machine Learning (ML)

In this article, we break down ML questions to ensure that you are familiar with the following:

• Elements of Machine Learning question
• ML solution submission
• Important terminologies

Elements of Machine Learning question

An ML question comprises of the following elements:

1. Question name - This is the title of the question.
2. Problem statement - A problem statement is a short description of the problem for which you are required to provide a solution or an answer.
3. Task - This section lists the specific tasks you need to complete in order to solve the problem. 
4. Dataset description - This section describes the provided dataset. It includes information about the files, the number of columns, and what each column represents. It may also explain any important details about the data that are needed to solve the problem.

Dataset

A dataset comprises:

• • Training dataset
  • Test dataset
  • Sample submission file
    

Training data set

A training data set is the data that you will use to train your models.

Test data set

A test data set is a data that you will use to predict your answer.

Sample submission file

A sample submission file consists of the format that you should follow while creating your submission file.

Note: All the variables in the test and training datasets are also described in the Data section. 

5. Evaluation metric - Your submission is evaluated based on many evaluation metrics, which depend on the type of problem. The evaluation metric that will be used to evaluate your submission is described in the problem.

Note: There are many evaluation metrics, such as Mean Absolute Error (MAE), F1 scoring, etc.

ML solution submission

You can submit multiple solutions for a machine learning question, and the best score will be taken as the final result.

New submission

You must upload the submission file and source code of the solution here.

Submit submission file

To submit your submission file, follow these steps:

1. In the Upload File section, click Upload File.
2. Navigate to the folder where you have saved your submission file.
3. Select the file and click Open.

Submit source code

To submit your source code, follow these steps:

1. In the Upload Source Code section, click Upload File.
2. Navigate to the folder where you have saved your source code.
3. Select the file and click Open.
4. In the Your Answer text box, write additional information (if any) related to your answer.

To submit your solution, click Submit. You have successfully submitted your solution.

All submissions

You can also refresh your submissions by clicking Refresh All Submissions List. Your last submission will appear on the top after you refresh this list.

Important terminologies

Test cases

The types of test cases are as follows:

Offline test cases

It contains all the test data based on which the final evaluation (after the challenge gets over) is done.

Online test cases

It contains some part of the test data based on which you are evaluated instantly after you submit your solution.

Submission

You must submit your prediction or submission file in the format as described in the sample submission file.

There are two types of submissions that you should make:

Online submissions

When you submit your solution, it is run against online test cases. You can make multiple submissions. These submissions will be evaluated instantly based on which leaderboard is updated.

Offline submissions

When you submit your final solution, it is run against offline test cases. Once you make this submission, you cannot edit it.