Combining Semantic Web technologies and Computational Theory of Perceptions for text generation in financial analysis

The development of Internet technologies makes it possible to obtain data in near real time about the financial state of companies. Moreover, tools such as XBRL have been developed to deal with the automatic generation of business reports. However the available tools are not suitable to support the current tendency towards the so called, Continuous Reporting. Here, for a specific purpose, the wealth of available information should be summarized in short reports where the relevant information is remarked while the irrelevant remains hidden. Using a combination of Zadeh´s Computational Theory of Perceptions and Semantic Web techniques, we propose a first approach to a computational framework to generate short reports that consist of few but meaningful Natural Language sentences. These sentences are designed to solve the user needs in a specific context of meaning defined with the help of an ontology. A practical demonstration of the computational framework is included.