Cancel
Showing results for 
Search instead for 
Did you mean: 

Confusions about MindSphere Prediction and Outlier Function

Experimenter
Experimenter

Dear All,

I have some confusion about MindSphere Prediction and Outlier function.Hopefully anybody can give me some instructions.

1. As for MindSphere Trend Prediction, is it possible to predict one Sine wave if I input one group of Sine wave data for its data source?

2.As for Outlier function, the outlier only means the function"> set value",no function "< set value"?

 

With Best Regards

Magica

3 REPLIES

Re: Confusions about MindSphere Prediction and Outlier Function

Siemens Valued Contributor Siemens Valued Contributor
Siemens Valued Contributor

@Magica here are a couple responses for your questions.

 

1. The trend prediction is based on existing asset/machine time series data - you could upload a sine wave data set to your existing asset and then the prediction API could use that as input to predict future values (a sine wave based on the historical data).

 

2.  The outlier detection API is currently set for only positive values only, unfortunately there is not yet a negative threshold available.

 

Hope this helps.

Mike 

Re: Confusions about MindSphere Prediction and Outlier Function

Experimenter
Experimenter

@Mbossom Hi Mike,

 

Thanks for your response. So as you said, the Prediction API could generate one sine wave if the historical data uploaeded to Clould is also Sine Wave? it won't be only one linear line(Increasing or Decreasing)? Any document or any Examples releases for Prediction elaboration?

 

Thanks 

Magica

Re: Confusions about MindSphere Prediction and Outlier Function

Experimenter
Experimenter

Hello!

 

Outlier checks absolute difference between two consecutive points and compares it against threshold, there are no min/max limit.

TrendPredictor does linear regression.