I want perform text categorization using topic models. In order to implement the existing systems to learn how they work, is it ok to use Libraries for this purpose or do I need to code the whole framework, so that the results could closely match.
thank you
Actually, There are so many opensource code existing on the website. If you are familiar with java, you can read code from here.https://github.com/hankcs/LDA4j or http://blog.csdn.net/yangliuy/article/details/8457329. Although it may seem a little harder, you can change the code according to your research. What's more, you can also use some powerful toolbox such as TMThttp://nlp.stanford.edu/software/tmt/tmt-0.4/, and http://psiexp.ss.uci.edu/research/programs_data/toolbox.htm. Lastly, the packages lda from R platform http://mirrors.opencas.cn/cran/ is also a good choice for you.
Thank you Ziheng,
I am starting to work on the research problem I proposed in this area. I have found help codes in Java (which as you pointed, could be changed and re-used). So should I go with Java to implement the problem. I will be using network analysis for text categorization.
About existing tools and Libraries (as you have generously mentioned few), is it Ok to use them, as the tools and libraries I use will also contribute towards the efficiency of the solution that i propose.
Thank you,
There is a "WEKA" which is a big library designed and found for data and text mining techniques to be used either with Java platform or as independent tool.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India