Chatbots have proven success in handling massive surges in call volume that the contact centers are witnessing globally during this pandemic. The reasons why chatbot technologies are being used is clear: 

• Promises consistency in all customer interactions
• Deploys immediate information irrespective of the time
• Enables instant scalability to meet demand spikes
• Handles huge query spikes in a timely and consistent manner.

A series of other companies have also started implementing pandemic-specific chatbots for answering questions coming in from their concerned customers.

Task

Your task is to predict whether a user is infected with the coronavirus based on the conversations between the user and the chatbot.

Data description

The dataset folder contains the following files:

• The trainConversation folder: Contains 4900 .json files that represent the conversations between each user and the chatbot
• The testConversation folder: Contains 2100 .json files that represent the conversations between each user and the chatbot
• train.csv: 4900 x 2
• test.csv: 2100 x 1
• sample_submission.csv: 5 x 2

The columns provided in the dataset are as follows:

Evaluation metric

score = 100 x metrics.f1_score(actual, predicted)

Result submission guidelines

• The index is userID and the target is the Prediction column. 
• The submission file must be submitted in .csv format only.
• The size of this submission file must be 2100 x 2.

Note: Ensure that your submission file contains the following:

• Correct index values as per the test file
• Correct names of columns as provided in the sample_submission.csv file