Since the 2008 sub-prime mortgage crisis and the global economic meltdown of 2009, investors have understood the value of historical data affecting a stock during investment. The general skeptical tendency of an investor increased in cases of mortgage bonds and general equities. Since then, the investment banks and financial institutions have increased their reliability on the available stock market data and finding correlations from the history of a stock to determine future trends and analyze current trends before making a decisions on the action to be performed on a particular stock. Using data mining techniques like clustering and association rules has helped quantitative analysts to determine correlations between the trend a stock follows and the reaction of a customer to the change in trends. This proves to be crucial in making recommendations to a customer on a particular financial product. The proposed research paper and eventual system, aims at providing a platform for data entry of large data sets of stock data, a set of input variables (which can be two or more) upon which various clustering and associative algorithmic techniques are implemented, to produce a result, similar to the result of a recommendation system, which predicts the next possible choice of a user based on historical ratings. The difference being that the ratings will be replaced by certain data points which can be system or user fed, and a correlation will be drawn between two or more variables with respect to these specific data points. The following paper lists the principles to be followed, the architecture of the system and the algorithmic techniques to be implemented.

Mortgage crisis stock market data correlations clustering recommendation system data mining chart pattern recognition classification large data sets