How to Use TimeSeries Model(created in analytic builder) with Thing to Predict next value in Series.

[MayankSharma20]

we have Time Series model created in Analytics Builder, Now I have to use it for prediction future values in series.

How can I use it in a Thing Object as service to predict future values and use it in Mashup UI for RealTime and Bulk (Batch of Time).

Stuck while doing trial-error- 

1. I was trying to create a service in Thing Object.

2. I have created Data shape and Properties for this Thing.

3. Created Analysis Provider, Event, and model Publish.

Actually i was following this article and try to replicate for Time Series.

https://community.ptc.com/t5/IoT-Tech-Tips/Analytics-Prediction-Methods-Mashup/m-p/570921 

Enclosed resource, errors, and Analysis Mapping(not able to see field for mapping).

Please suggest some thing service snippet for Time Series model Utilization.

[cmorfin]

Hi @MayankSharma20 

I noticed that you already have a case opened with Technical Support on this topic.

This is the right thing to do.

My colleague, Mohit, from Technical Support will contact you on Monday on this.

Here is a summary from what I could see.

- you have published a model into Analytics Manager.

- In this case you do not need to create any service, the Analysis Event will take care of triggering the jobs

- you only need to have a Thing that receive input data

- you seem to be confused with the input datashape but it is as expected. Your model has been created with all the field excluded except the goal Clothing. So the only input is Clothing which happen to also be the goal.
This means that the Clothing field will be predicted from previous value of itself. If this is not what you expect, you should create the model differently. If this is what you expect, then simply map a property of the corresponding thing in the Analysis Event to the Clothing field.

- This article https://www.ptc.com/en/support/article/CS293493 should be of interest too.

Mohit will be able to give you more details if needed through the Technical Support case.

Kind regards

Christophe

[MayankSharma20]

Thanks for your valuable response..!

As attached dataset, I can select any goal for prediction.

Rec_goods

Clothing

Footwear

Chemist

Main thing

1. I am confused in datashape part and the way to read/consume model prediction using Thing.

2. Even i am not able to see fields while mapping data in Analysis Event.

I was looking the way to mapping, and the way(code snippet) to utilized the Time Series prediction model using Thing. 

Regards,

Mayank

[MayankSharma20]

Hi,

Please check our use case defined below:

1. We have 3 years of sales data (monthly basis) which has features like Sales(Float), Entity_ID(Integer), Time(Integer). And need to predict next 1 year Sales.

2.  We have to do this prediction in a Batch, And wanted to automate this by uploading next year Sales Data and Store prediction results in datatable or database which can be utilize by model for future training and predictions.

3. We need to automate model training and prediction.

@steps in detailed:

1. Upload the historic data.

2. Create a new model with this historic data.

3. Upload new data – the one to be scored.

4. Perform a prediction job to score those new data and store in Database or DB Table.

5. Retrieve the prediction job result.

6. Automate prediction model by utilizing store data for future training and prediction.

If you know anything please let me know.

Regards,

Mayank

[mgoel · Accepted solution]

Hi Mayank,

After following this Link, your issue has been resolved over PTC support case. Please mark this thread accept as solution for benefit of others.

Regards,

Mohit

[mgoel]

Hi Mayank, 

Since this case has been resolved with PTC technical support, could you please mark this thread as resolved to for reference of other guys.

Reference link:  https://community.ptc.com/t5/IoT-Tech-Tips/How-to-work-with-ordinal-and-categorical-data-in-ThingWorx/m-p/542406 

Regards,

Mohit

[slangley]

Hi @MayankSharma20.

If you feel your question has been answered, please mark the appropriate post as the Accepted Solution for the benefit of others with the same question.

Thank you for participating in our community!

Regards.

--Sharon