Financial Market Web Mining with the Software Agent PISA

The World Wide Web (WWW) contains millions of hypertext pages that present nearly all kinds of information. Specific effort is required to use this information as input to subsequent computations or to aggregate values in web pages. To do so, data from webpages are extracted and stored on computers either in text files or databases. These manipulations can be better attained by using an autonomous software program instead of human personal. Here, we present a platform independent software agent called PISA (Partially Intelligent Software Agent) that extracts financial data autonomously from webpages and stores them on a local computer. Data quality has highest significance. PISA generates time series with user defined denseness. These time series have adequate quality for financial market analyses, e. g. forecasts with neural networks.