Prediction of cost based on the historical data

Hi Guys,

I want to build a model which will predict the coming month costing based on the historical data which have following inputs
Period : Month wise data( number)
Cost Center : Pure Number( eg. 1199234) as well as text and numeric ( eg NA233447757)
Organization : String type
Service : Various functions in the Organization( String type)

Kindly provide insight to build the predictive model based on above mentioned inputs.


This is a regression problem. You could have a look at the workflows on the knime hub regarding prediction workflows.
e.g. sth like this



@singhsukh12 you will have to decide if this is a problem that indeed can be formulated as a regression task, this will depend on the relationship between your variables and he target.

A collection of regression algorithms can be found here

Otherwise this might be a time series problem if the costs are seasonal.


Projects like this should really be considered a financial forecasting model and not just a mathematical calculation based on historical cost data. I don’t recommend approaching problems like this in a vacuum or from a purely analytic perspective. This data should be generated during a good budgeting process that takes into account budget initiatives, fixed Vs variable costs, historical data, inflationary predictions, business plans, etc.

I am not discounting the regression analysis recommendations, as that is a valid process approach to the question. Just want to add a warning against using purely statistical forecasts in budgeting or management reports for strategic decision making. I hope this didn’t come across too preachy… I am not assuming that the OP’s approach in particular would have been lacking financial discipline in its final implementation. This is just a common mistake in financial management, and I thought it might be a helpful post for the wider community.


agree with iCFO, forecasting like this should be considered financial forecasting models, not just mathematical calculations based on historical cost data. financial forecasting models should be dynamic and responsive to changes in the business environment, and able to simulate different scenarios to support informed decision-making … just my two cents.

1 Like

Thanks for sharing your solution . Just one more Point as mine data is monthly basis and i also want Period column to be as variable so that regression learner will also take this column into consideration while building the model to give me the output. How that can be possible .

Yes data is coming on monthly basis based on which prediction to be done for next month or coming month.
Any solution for the same where is time series option can be included.
Thanks for your response for earlier post .

@singhsukh12 you should think about what @iCFO said. The ‘answer’ will have to be in your data. Otherwise you will not be able to have a model.

And you also have to take into consideration let’s call them (business) ‘mechanisms’ that might influence your values like taxes, competition, seasons … this will require domain knowledge.

Then if you decide that a regression or time series model is indeed what you want to use you will have to build a data set that best represents your challenge on historical data (with the addition of the target) and that you also can reliably reproduce for future predictions. And then you will have to think what you will do with a ‘prediction’ - what would be a decision you would take if the model comes back with a certain number. This answer might also shape you model strategies (choice of metrics eg.).

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.