MSc Thesis, University of Tuzla, Bosnia.

**ISSN/ISBN:** Not available at this time.
**DOI:** Not available at this time.

Note - this is a foreign language paper: SRP

- For online information, click here.

**Abstract:** Benford’s law describes the probability for leading digits of numbers. It is named after Frank Albert Benford (1938) who stated a mathematical model of this probability. Before him, the
same observation was made by Simon Newcomb. This law has changed the usual assumption of equal probability of each digit on each position. A main goal of the text is to present posibilities, advantages and disadvantages of Benford’s law in
reinforcement learning methods. One particular goal is to analyze the influence of values generated by Benford’s law and defined in patent applications by Fletcher Lu and Efrim Boritz,
in which they formulated an Adaptive Benford’s method. This value is used as reward for statest in reinforcement learning. In this text the usage of quotients of values for three and two
leading digits is proposed.

**Bibtex:**

```
@mastersThesis{,
AUTHOR = {Jasak, Zoran},
TITLE = {Benfordov zakon i reinforcement učenje (Benford's Law and reinforcment learning)},
SCHOOL = {University of Tuzla},
YEAR = {2010},
ADDRESS ={Bosnia},
}
```

**Reference Type:** Thesis

**Subject Area(s):** Computer Science, General Interest, Statistics