Big data-based stock market inner-screen transaction behavior identification method

The invention discloses a big data-based stock market inner-screen transaction behavior identification method. The method comprises the following steps of obtaining company related data; calculating the increment of the stock trading volume mean value; if the increment of the stock trading volume mean value is greater than a threshold, calculating the increment of the average value of all stock trade volumes of the industry; if the increment of the stock trading volume mean value is less than a threshold, inquiring whether to publish announcements in the earlier stage; calculating the fluctuation amplitude of the stock closing price caused by external announcement; if the fluctuation threshold is not exceeded, calculating whether the weighted average net asset yield of the listed company is lower than the weighted average net asset yield; and if so, acquiring an external announcement of a later listed company, judging whether the external announcement is a good announcement; if so, calculating the maximum ris