Comparing with the experimental results purely based on P, the mean predicting errors forecasted by P and W were obviously reduced.
The addition of W thereof enhanced the forecasting performance. In theory, the investigation provides an insightful understanding of the functions of financial information time series from a new angle, and it can be theoretically treated as a new and promising way to probe into complex problems in security
markets.
As more and more ordinary people around the world share
information on the web, the ‘‘long-tail’’ formed by these ‘‘grass
roots’’ has a noticeable effect on the market. The rapid responses
of computer processing of financial information can help financial
managers quickly find the aggregated effect of opinions from
these ‘‘grass roots,’’ as well as other enormous web financial
information. The bridge we built in this paper between P and W
containing the views of ‘‘grass roots’’ offers a novel clue to P from
the W side for financial analysts, and enables them to rapidly