This system is developed as a part of academic main project. This system can be used to predict the authorship of an unknown document. You need to train the system before the prediction. Training means you need to give author names and some books of the author to the system. We got 80.95% output in overall testing and 92% result in Binary testing. A video demonstration of working can be seen in this Working video. System is fully written in python. You need some extra python modules to make the system work. You can find the detailed instructions to install these modules under the 'Installation instructions' sub- heading. You need to copy the 'SourceCode' folder to your linux system and use the below command to run the system
user@system :~$ python 1.py
You can find DataSet containing some author's documeents in a folder called DataSet. We used 'git' for the version control purpose. So you can access all of the source code history by make use of the git command. This Source code is free to use for study purpose only. Any commercial usage of this source code with out the written permission from the authors is strictly illegal. For any doubts feel free to drop a mail to this address [email protected] or make a phone call to +919961719738
***************
Installation Instructions
-----------------------------
Type the following command in the terminal to install the following:
• Python
user@system :~$ sudo apt-get install python
• WXPython installation
user@system :~$ sudo apt-get install python
• nltk toolkit installation
user@system :~$ sudo apt-get install python-nltk
• numpy installation
user@system :~$ sudo apt-get install python-numpy
• matplotlib installation
user@system :~$ sudo apt-get install python-matplotlib
• scikit-learn installation
--> Download source of scikit-learn source from the following link.
https://github.com/scikit-learn/scikit-learn
--> Extract it to home folder.
--> Open terminal.
user@system :~$ cd ~/sklearn/
user@system :~$ python setup.py build
user@system :~$ python setup.py install
**** Thank You ****