Publication detail

System for intelligent investment portfolio making

BUDÍK, J. DOSTÁL, P.

Original Title

System for intelligent investment portfolio making

Czech Title

Systém pro tvorbu investičního portfolia

English Title

System for intelligent investment portfolio making

Type

journal article

Language

en

Original Abstract

The paper describes system for building intelligent investment portfolio based on a profile of a trader using these rules. The fact, that 60% of successful trading is a psychological part, is very important. The main idea is based on the fact that successful traders have to know about his features as a tolerance of risk, level of patience and other personal features. System works with three basic profiles of trader, in which is assigned trader after answering a few questions. If the profile of trader is known, the next step is to search an optimal investment portfolio by combining fuzzy logic and data mining methods to find rules for decision making. The main concept in searching for a statistical edge is based on the fact, that markets have everyday tendency to move in one direction and it doesn't matter which side is it (up-move or down-move). Important is, that there is usually move to one direction. Data mining methods will find time periods with big probability for a price move to one side. We achieve this by comparing current price to reference price (Open, High, Low, and Close of n-previous price bars). Founded edge should be wedged by some fundaments like market opening, market closing etc. That is another view for entry point other than looking for combination in technical indicators. Diversification is realized through two or more non-correlated markets. More important rule than entry point is exit point and that is the thing, where is difference between trader profiles, which are set up in the beginning of the process using fuzzy logic. Decision making rules are set from January 2007 to December 2008 and used from January 2009 to January 2010, so we can see robustness of these rules through the financial crisis.

Czech abstract

Článek pojednává o systému pro tvorbu investičního portfólia. Je uvedena teorie, vystvěn model a uvedena případová studie. Metoda je založena na statistických metodách.

English abstract

The paper describes system for building intelligent investment portfolio based on a profile of a trader using these rules. The fact, that 60% of successful trading is a psychological part, is very important. The main idea is based on the fact that successful traders have to know about his features as a tolerance of risk, level of patience and other personal features. System works with three basic profiles of trader, in which is assigned trader after answering a few questions. If the profile of trader is known, the next step is to search an optimal investment portfolio by combining fuzzy logic and data mining methods to find rules for decision making. The main concept in searching for a statistical edge is based on the fact, that markets have everyday tendency to move in one direction and it doesn't matter which side is it (up-move or down-move). Important is, that there is usually move to one direction. Data mining methods will find time periods with big probability for a price move to one side. We achieve this by comparing current price to reference price (Open, High, Low, and Close of n-previous price bars). Founded edge should be wedged by some fundaments like market opening, market closing etc. That is another view for entry point other than looking for combination in technical indicators. Diversification is realized through two or more non-correlated markets. More important rule than entry point is exit point and that is the thing, where is difference between trader profiles, which are set up in the beginning of the process using fuzzy logic. Decision making rules are set from January 2007 to December 2008 and used from January 2009 to January 2010, so we can see robustness of these rules through the financial crisis.

Keywords

Portfolio, system, datamining, time series, decision making

RIV year

2010

Released

23.06.2010

Publisher

International Institute of Forecasting

Location

USA

Pages from

1

Pages to

6

Pages count

6