Is regression a good method for prediction when a variable is dependent on multiple attributes each having seasonal variations
For simple or multiple regression, you can’t directly conduct regression method for your data, since one of the assumption for simple/multiple regression is independent variable should not have multicolinearity among them. Maybe you can use robust regrssion or use time series method (ARIMA) that can address your data.
It is quite common to perform seasonal adjustment of data where there is a strong seasonal trend, then use the seasonally-adjusted data in further analyses. Many textbooks and websites describe techniques for doing such adjustments. With daily numbers you must take into account both day of the week AND season, and this can get a little tricky. For example December 2011 was odd because by the last pre-Christmas Saturday, which is usually a big shopping day, many of us had already finished buying gifts since Christmas fell on a Sunday. Anyway, look up how to do such adjustments, and you can use the adjusted data as the inputs for linear regression.
But in my data there is no collinearity and whereever it exists i can reduce those attributes during dimensionality reduction process
FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!