# Context This is probably the dumbest dataset on Kaggle. The whole point is, however, to provide a common dataset for linear regression. Although such a dataset can easily be generated in Excel with random numbers, results would not be comparable. # Content The training dataset is a CSV file with 700 data pairs (x,y). The x-values are numbers between 0 and 100. The corresponding y-values have been generated using the Excel function NORMINV(RAND(), x, 3). Consequently, the best estimate for y should be x. The test dataset is a CSV file with 300 data pairs. # Acknowledgements Thank you, Dan Bricklin and Bob Frankston for inventing the first spreadsheet. # Inspiration I hope this dataset will encourage all newbies to enter the world of machine learning, possibly starting with a simple linear regression. # Data license Obviously, data is free.