Finance is one of the many fields which has received large benefits in the last 20 years from the availability of high technological tools and their interconnection. Markets can now be accessed very quickly even by non-professional investors and market data are continuously stored by computers which are powerful enough to handle these huge amounts of data and perform real-time analysis on them. Therefore investors have nowadays access to a vast amount of information, provided by agency press news providers, historical archives, government agencies and private organisations which collect, organise, process and distribute different types of data. They are connected in real-time through the Internet to the news providers, drastically reducing the lag between when news happens and when it is assimilated by the operator. The cost of this information has decreased rapidly, making it accessible even by small and private investors.
The very large amount of information to which financial operators have access is therefore of a varying degree of quality. Quantitative information consists of information which can be expressed by numbers and which can therefore be automatically used, without any further elaboration, for analysis by investors or by computer programs. Qualitative information is, on the other hand, information which may not be expressed in numeric terms.
Financial tools are instruments used to perform a selection of qualitative and quantitative data and to produce new inferred data. We can define them as tools that are used to support the decision making process of the investors, by summarising, selecting or producing some analysis of the original source data or suggest some operative decisions. They are, therefore, instruments used to perform economical analysis of the stock market. Many different kinds of financial tools have been developed. Our research work is on tools used in decision-making processes dealing with the trading of securities issued by companies and quoted in the stock exchange, but which can be extended to other financial markets.
Conventional tools are based on mathematical, statistical and probabilistic algorithms. These techniques are deeply studied, highly optimised and effective on particular sets of data and have been experimented for a long time. Conventional tools we work on are based on statistics, which perform analysis and calculations on data consisting of historical time series of share prices and on heuristics, which try to explain the financial events without an underlying economical or mathematical relation, but basing on experimentation, trial-and-error algorithms and knowledge available within the financial community.
Innovative tools we work on have instead been introduced in the last 20 years and are mainly based on techniques borrowed from Artificial Intelligence, such as neural networks, a computational model for information processing based on a connectionist approach and expert systems, computer programs which possess some specific knowledge which usually belongs to human experts.