Using Intelligence Predictive Model

[Tiger1]

I am trying to use the intelligence Predictive Model and i am not getting any data from it. It does not predict anything.

[Arthur_Rallu]

Hi Tiger,

When you add a new record in that table with new values for all these months, are you not seeing any new value in the ‘Total Prediction’ column?

[Tiger1]

Oh. This is based on adding new data? I put this table together based on historical data - I won’t be adding or changing it. But I wanted the prediction model to predict next year’s revenue.

Shane

[Tiger1]

How do i make the model predict without having to add new data? I created a table that has four years of data. From this table i want the prediction model to predict next years revenue. So, I am not adding any new info/data. Can this work?

[Arthur_Rallu]

So you have four years of past revenue and revenue only, and you want to forecast next year’s revenue without any input. Is that correct?

[Tiger1]

Yes.

[Arthur_Rallu]

Currently, this is beyond the scope of our predictive models.

Our predictive models can predict a new outcome from new input data, based on what is already known. They can be based on past data (past inputs + past outcome), but it could be other data as well. This is different from forecasting where you make a prediction based not on new inputs but rather on what has happened over a certain period of time in the past.

[Tiger1]

Ok. I have the table (revenue from past) but to get that table built - i had to use inputs (items out X rate of the item (per month)). So i do have inputs (items out and the rate of the item (how much we charge)), but the table i built is just showing the revenue (so it only shows the actual revenue - not how it got the revenue). But you are saying, if i have the inputs (items out and Rate of the item) i can predict? Also, are there others predicting (Forecasting) revenue based on past?

[Arthur_Rallu]

If you know how many items you’ll be selling next year, you can compute next year’s revenue by applying the same formula (‘items out’ X ‘rate of the item’) and you don’t need a predictive model for that.

Now, how do you know the ‘items out’ for next year? It seems to me you would still have the same problem, you still need to forecast the ‘items out’.