/U-919-SupportPredictor

EMC underground Alexa Skills challenge project; the skill predicts an upcoming service request for given frame.

Primary LanguageJavaScript

Alexa Challenge

A simple Amazon Echo project to participate in the EMC Underground Challenge

Support Predictor

The solution's goal is to predict whether a support case will be open for a specific EMC array. The prediction is based ML (Machine Learning) inside an AWS lambda function using an existing historical data set.

Logic:

  • Initialize ML NN (neural network)
  • Load a training set for a specified frame
  • Ask user for a date
  • Support Predictor tells the chances for a new support request being open for a frame

Sample dialog

  • (User) "Alexa, use Support Predictor"
  • (Alexa) "Welcome to the Dell EMC Support Predictor. I have data for two frames: 1234 and 5678. Please tell me which one to use?"
  • (User) "Help!"
  • (Alexa) "Please chose one of the ZZ frames for a prediction about an upcoming support request. You may also ask to list the frames."
  • (User) "List frames please"
  • (Alexa) "I have data for ZZ frames. And they are: 1234,2345,3456,4567. Please choose a frame."
  • (User) "Frame 1234"
  • (Alexa) "Building the neural network for frame 1234. Built successfully! Which day should I predict for?"
  • (User) "Today" || "Tomorrow" || "A month from now" || "day after tomorrow"
  • (Alexa) "There is a XX% chance that there is an issue on DATE!"

Development

The project is using NODE.JS. The ML component is the "synaptic" library; also planning to use AWS S3 for the lists.

cd src && npm install synaptic aws-sdk