Credit scoring techniques application in investment decision making
| Authors | |
|---|---|
| Year of publication | 2014 |
| Type | Article in Proceedings |
| Conference | New Economic Challenges – 5th International PhD Student Conference |
| MU Faculty or unit | |
| Citation | |
| Field | Management and administrative |
| Keywords | logistic regression; technical analysis; exponential moving averages; automated trading |
| Description | Algorithmic trading is achieving leading position in contemporary development of financial markets. Most strategies of algorithmic trading are still based on technical analysis. Basic technical analysis is following the rule that chosen strategy should be easy understandable and applicable in order to simplify investor’s decision making. This paper analyzes more sophisticated methods to combine the signals from several indicators to design investing strategy, which would be profitable in the long run. Econometrical and credit scoring methods are used to determine whether chosen indicators are relevant to the designed strategy. Explaining variables are time series of dummy variables indicating whether the indicator (exponential moving average) is suggesting submitting buy order or not. Lags of these variables are also included, so that time relations of indicators can be tested and whether indicators confirm each other. A dependent variable is the successful trade (situation when the positive price movement is greater than the spread and commissions). Logistic regression, which is widely used in credit scoring, is one of most appropriate tool for this research. This paper shows that chosen methods can be used to profitable and efficient model, which outmatches indicators themselves, which was proven on the simulated trading of exchange rates of several currencies. |
| Related projects: |