Lefeber, F. (2013) Random Writing And Authorship. Bachelor's Thesis, Mathematics.
|
Text
MainThesis.pdf - Published Version Download (493kB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
AkkoordWit.pdf - Other Restricted to Registered users only Download (40kB) |
Abstract
The question of authorship of texts has already been investigated by several scientists. For example, in the Journal Of The American Statistical Association, authorhip was determined with the help of context free words that were called function words.~cite{MW} In this paper a different method for determining the authorship of a text will be explored and analysed. The verdict of the origin of a text will be based on likelihood ratio tests between candidate authors. The loglikelihoods that are necessary are unknown, but they will be estimated. These estimates are derived from the probability that the candidates write the text using a stochastic model. This model is designed to simulate the writing style of an author. It contains a Markov Chain based on n-grams; transitions between n-tuples of words in the text. It will use the maximum likelihood estimator to assign probabilities to each transition. For some analysis, the topic of ergodicity will be briefly covered, along with the corresponding conditions. In order to test whether the method is suitable for authorship testing, we built the required functions in MATLAB. There is also a section devoted to the way the method was implemented in this programming language.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Degree programme: | Mathematics |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 15 Feb 2018 07:53 |
| Last Modified: | 15 Feb 2018 07:53 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/11081 |
Actions (login required)
 |
View Item |