Recognizing in advance the possibility that a financial event will take place is very important for investors, to allow them to make the best decision. For this reason several studies have been carried out and many automatic tools are available to support decision making. This thesis will present Mercurio, a system that supports the decision making process of financial investors through the automatic extraction of information from financial news and the analysis of financial data retrieved from the Web. Starting from the formalization of the knowledge of an expert in financial journalism, Mercurio can extract information from financial newspapers and use it to identify relevant financial events. In parallel it identifies relevant events related to the stock market by an automatic analysis of financial indexes. Lastly, by using sequential pattern mining algorithms, Mercurio builds a model to predict the happening of exceptional events given their past occurrences and relationships with other events. This allows investors that use the system to be warned in advance about the possible financial event so they can choose the best time to place themselves in the market. The aim of this research is to produce an innovative system that supports the decision-making process by combining the analysis of the financial indexes with a semantic analysis based on the reasoning of a financial journalism domain expert.
Riconoscere anticipatamente la possibilità che si verifichi un evento straordinario nell’andamento del mercato finanziario è fondamentale per gli investitori, per poter prendere la miglior decisione possibile. A questo proposito vari studi sono stati condotti e molti strumenti automatici in grado di fornire supporto alle decisioni sono disponibili. In questa tesi viene presentato Mercurio, un sistema che supporta il processo decisionale degli investitori, mediante l’estrazione di informazioni da articoli di carattere finanziario e l’analisi di dati finanziari ottenuti dal Web. Partendo dalla formalizzazione della conoscenza di un esperto in giornalismo finanziario, Mercurio permette di estrarre dal contenuto degli articoli dei quotidiani finanziari le informazioni necessarie ad indentificare eventi finanziari rilevanti. Parallelamente, identifica attraverso un’analisi automatica degli indici finanziari gli eventi maggiormente rilevanti nell’andamento del mercato azionario. Infine, utilizzando algoritmi di sequential pattern mining, Mercurio costruisce un modello predittivo basato sulle passate occorrenze e sulla relazione tra gli eventi individuati. Tale modello permette agli investitori che utilizzano il sistema di essere allertati anticipatamente riguardo il possibile evento finanziario consentendo loro di posizionarsi nel mercato al momento più opportuno. Lo scopo di questa ricerca è quello di produrre un innovativo sistema di supporto alle decisioni che affianchi all’analisi dell’andamento degli indici finanziari un’altra analisi, di tipo semantico, basata sui ragionamenti di un esperto del dominio giornalistico finanziario.
Detection of significant financial events from newspaper articles to support investors' decision-making
BRAMBILLA, MARCO
2015/2016
Abstract
Recognizing in advance the possibility that a financial event will take place is very important for investors, to allow them to make the best decision. For this reason several studies have been carried out and many automatic tools are available to support decision making. This thesis will present Mercurio, a system that supports the decision making process of financial investors through the automatic extraction of information from financial news and the analysis of financial data retrieved from the Web. Starting from the formalization of the knowledge of an expert in financial journalism, Mercurio can extract information from financial newspapers and use it to identify relevant financial events. In parallel it identifies relevant events related to the stock market by an automatic analysis of financial indexes. Lastly, by using sequential pattern mining algorithms, Mercurio builds a model to predict the happening of exceptional events given their past occurrences and relationships with other events. This allows investors that use the system to be warned in advance about the possible financial event so they can choose the best time to place themselves in the market. The aim of this research is to produce an innovative system that supports the decision-making process by combining the analysis of the financial indexes with a semantic analysis based on the reasoning of a financial journalism domain expert.File | Dimensione | Formato | |
---|---|---|---|
2017_04_Brambilla.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Testo della tesi
Dimensione
2.69 MB
Formato
Adobe PDF
|
2.69 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/133921