Data Analysis and Finding Relationship

I have a set of files with 250+ columns of data and outputs from that data (less than 10 columns) which are calculated based on these inputs. Given that not all outputs rely on all inputs, and the relationship between the columns they are related to may vary (e.g. be logarithmic) I need mathematical analysis (possibly regression analysis) to find the relationship between the inputs and outputs. The way of calculating the relationship has to be found. The deliverable of the job would be the mathematical formula/algorithm and an excel sheet calculating the output from input data. I have numerous similar files on which this can be tested to verify the output. RESTATING THE OPPORTUNITY AGAIN: I have a range of counters (250+ counters). I also have some related measurements along with these counters. Using the values of these counters, and some measurements, I need to arrive at some outputs. I have the outputs (arrived for by an unknown algorithm) for known values as well. And I have enough data. What do I need? I want formulae to compute each of the outputs, based on the values of the counters and measurements. Basically I am looking for someone to reverse engineer the algorithms/formulae behind arriving at the outputs given the inputs (counters and measurements) I would need a mathematical formula as well as an excel formula to calculate the outputs Hints: - Not all counters are relevant to all outputs - I believe there may be some sort of logarithmic regression involved - I have test files to validate your algorithm and can provide more test data if required This would be a good mathematical investigation project In the attached files, the output columns B and C are the calculated outputs, everything else, specially the inputs files are all inputs. The entries are timestamped and each corresponds. You could merge both and then work on that.