Stock price trend prediction method and system

The invention discloses a stock price trend prediction method and systemThe method comprises the following steps: setting a data set of a stock as a two-dimensional set D, and enabling a horizontal axis to represent the date attribute and other attributes of the data set, and recording the date attribute and other attributes as {Data, a1, a2,..., ak}; defining a time correlation degree of the setD; defining a split node set Tk of other attributes ak, and defining a time correlation gain of a split node t; defining an information gain of the split node t; combining the time correlation degreegain TRG (D, t) and the information gain EG (D, t) to obtain a decision tree split node standard; training the training set through the above steps to obtain a decision tree model, dividing the training set D * into a plurality of subsets according to dates, the set of the subsets being C [gamma] = {C1, C2,..., C [omega]}; and setting a formula for judging left and right sub-tree selection basis of the prediction set at th