Skip to Content

Use Service Ticket Intelligence and Postman to Get Solution Recommendations

test
0 %
Use Service Ticket Intelligence and Postman to Get Solution Recommendations
Details

Use Service Ticket Intelligence and Postman to Get Solution Recommendations

July 9, 2020
Created by
July 9, 2020
Train a machine learning model based on historical service requests in order to get solution recommendations for new requests.

You will learn

  • How to upload training data and train a machine learning model to get solution recommendations for new service requests.
  • How to get solution recommendations for new service requests.

Service Ticket Intelligence provides machine learning functionalities to effectively deal with service requests. Thereby, the service offers two main functionalities:

  1. The service is able to analyze the unstructured information in service requests and classify the requests into categories.

  2. The service is able to recommend solutions for service requests based on historical data.

Find more details on the Service Ticket Intelligence concepts here.


Step 1: Upload training data

Select the request Upload Recommendation Training File (random Q&A - small). The training data is already included in the Body of the request. Click Send to upload the training data. The dataset contains question-article related data and has the following information: subject, datasource and article_id.

Upload Training File

The service now automatically creates a new model and sets its status to NEW. That indicates that the model was just created and has not been trained yet. Both the ID of the new model, as well as its status, can be found in the response of the request.

Upload Training File Response
Log on to answer question
Step 2: Train model

Next, you can start training the model based on the data that was uploaded in the previous step.

Therefore, select the request Start model training and click Send. This will trigger the model training process.

Start Model Training

To observe the current status of your model, select the request Get model status and click Send.

Inspect Model Status

The response of the request includes all the details of your model along with its current status. When training is triggered, the status is PENDING_TRAINING which indicates that the model is queued for training. Once training has started, the status changes to IN_TRAINING. When training is done, the status changes to READY. Make sure to send the request Get model status every couple of minutes until training is completed. Training time varies depending on data size and infrastructure resource utilizations. In this example, it should take approximately 10 to 20 minutes to complete.

Model Status
Log on to answer question
Step 3: Activate model

To actually get solution recommendations for new service requests using the trained machine learning model, it is necessary to activate (deploy) the model.

Select the request Activate model and click Send. This will start the activation process of the model.

Activate Model

Use the request Get model status from the previous step again to observe the process. Once the status of the model changes to ACTIVE, the model is ready and activated.

Activate Model
Log on to answer question
Step 4: Get solution recommendations

Finally, the model can be used to get solution recommendations for new service requests.

Select the request Recommend Solution. Open the Body tab to see the service request message that will be sent to the service. Click Send to get solution recommendations for this service request.

The request and response look as follows:

Get Rec RequestsGet Rec Requests

This is the full prediction:

{
   "results": [
      {
         "detected_language": "en",
         "id": 2001,
         "recommendation": [
            {
               "score": 1.0,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "405639"
                  }
               ]
            },
            {
               "score": 1.0,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "405352"
                  }
               ]
            },
            {
               "score": 1.0,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "mindtouch"
                  },
                  {
                     "field": "article_id",
                     "value": "406961"
                  }
               ]
            },
            {
               "score": 1.0,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "mindtouch"
                  },
                  {
                     "field": "article_id",
                     "value": "405980"
                  }
               ]
            },
            {
               "score": 1.0,
               "solutions":[
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "407027"
                  }
               ]
            },
            {
               "score": 0.5,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "406000"
                  }
               ]
            },
            {
               "score": 0.5,
               "solutions":[
                  {
                     "field": "datasource",
                     "value": "mindtouch"
                  },
                  {
                     "field": "article_id",
                     "value": "405727"
                  }
               ]
            },
            {
               "score": 0.5,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "407386"
                  }
               ]
            },
            {
               "score": 0.5,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "407725"
                  }
               ]
            },
            {
               "score": 0.5,
               "solutions": [
                  {
                     "field": "datasource",
                     "value": "jam"
                  },
                  {
                     "field": "article_id",
                     "value": "405552"
                  }
               ]
            }
         ],
         "status": 0,
         "status_message": "ok"
      }
   ]
}    

You can also try out the other POST /Recommend requests.

Get Rec Requests

You have now successfully used Service Ticket Intelligence to get solution recommendations for new service requests.

Log on to answer question
Step 5: Test yourself
Choose the step that comes right after starting the recommendation model training.
×

Next Steps

Back to top