calculate a forecast on past data

[ARock1980]

Hello everyone

I want to calculate a FC based on historical data. I have a column with the FC value and a column with the actual value received. I would like to calculate the values and the deviation from the actual values of a FC based on the past values.

predictiv FC.xlsx (12.8 KB)

This theme is absolutely new for me so pls help me

Thanks

[rfeigel]

Is “FC value” forecast?

[ARock1980]

@rfeigel Iv got the Data from the past till 01.07.2024. You see the forecastet value “FC Quantity” and the real Quantity “effective Quantity” and the difference. From 01.08.2024 i only got the Forecast values. My idea is based on information on the past to get Forecast Values or here “effectiv Quantity calculeted”. I want to derive from the past values and the difference between values for “effective quantity calculated”

[hmfa]

Hi @ARock1980 .
Have you looked at this wf?
Maybe it can help you

Price Forecasting and Reporting Data App – josegallardo

Br

[ARock1980]

@hmfa Thanks for your answer. Im sorry but i dont know how to integrate my datset in this WF. I always get an error

[hmfa]

Hi @ARock1980.
You can test it by renaming your excel columns like this,

But I suppose you can’t have a good prediction with only 9 rows of historical effective quantity data. This example has 277 history data points.
Best Regards

[rfeigel]

Could you try explaining what you’re trying to do again? Please use consistent terms in your explanation. Also, I agree with @hmfa that you have very little data to work with.

[mlauber71]

@ARock1980 you can take a look at these threads and examples. Though I agree with @hmfa and @rfeigel you most likely have too few data points to make a real decision.

Can KNIME be used to show the employment and unemployment rate in the UK from the year 2008-2018, showing the prediction for the next year?

Not 100% sure what you mean by ‘rendering’. It is possible to make a prediction. I added a workflow that does that and also extended the existing ones slightly: [image] I also set up an example that just uses the numbers of the previous year and used the last year as test. But the results seem to be worse. You could check it out if you want:

Unsupervised machine learning (gradient boosting regression)

I set up an example that should look like your data (if not feel free to adapt and inform). We have two years of data (from one city) and data from January and want to predict the coefficient for Feb 2020. [image] 2019 serves as training and 2018 as test data. You might have more complex settings with cross-validation if you want. And also since you have the data you might provide coefficients of some more past months. Each line represents its own time series so to speak. Of course, such a se…

If you want to read more about methods and philosophy of predictions read the liked articles.