I am new to Data Mining. Am using SQL Server 2012. I have a few questions to ask
1. I created a time series algorithm model. Got the following error when processing it - "Processing the model failed because either the training data is empty or there were no significant attributes found." What could the problem be?
2. At some stage in creating the time series model, i also got an error to do with a column using a content type that is not suited to the time series algorithm. The column itself in sql was an nvarchar field. In SSAS it was flagged as of Text type and Discrete content but it kept giving the error. The field has text with characters in places like % signs in it. Is that the reason why, and how can i resolve this?
3. I have a set of financials of a company from 2005 to 2010 and i have been asked to do forecasts per items in those financials for the next 5 years. Is the time series algorithm the best to use in forecasting?
I created a mining structure with the following fields - Financial Year, Company, LineNo, Financial Description, DateKey and Amount. The time series key is on the Financial Year, and the Key is on Financial Description. The Amount field is a PredictOnly field. Is this ok or there is something i need to change?