[This article was first published on RTextTools: a machine learning library for text classification - Blog, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
RTextTools v1.3.5 addresses some key concerns that have been raised in recent months. Many of the algorithms used in RTextTools require that any new data presented to a trained classifier contain the same features as the original document-term matrix. Since this rarely (if ever) happens in the real world, I have added an originalMatrix parameter to the create_matrix() function that adjusts new document-term matrices to contain the same terms as the original training matrix. Although this is a rather quirky work-around, it enables users to save trained models and classify new data easily. Example scripts are available in the /inst/examples/ directory of the RTextTools source code.
Since its introduction at the 2011 Comparative Agendas Project Conference in Catania, Italy, the RTextTools team has refined the API and implemented a number of features. Some of these features include n-gram analysis, text labels, comprehensive analytics, and a streamlined interface. Our plan for the year ahead includes a major overhaul of the nine algorithms to facilitate low-memory ensemble classification. However, this goal involves more than just the RTextTools team; it requires the R machine learning community to strive for efficient supervised learning algorithms. Many R packages do not utilize compressed sparse matrices, and therefore are limited in their applications for large-N data-sets. Therefore, we aim to promote efficient practices by package developers and write several implementations of our own to push the capabilities of statistical computing in R.
Thank you for all your feedback and support as we look forward to another productive year in 2012!
Since its introduction at the 2011 Comparative Agendas Project Conference in Catania, Italy, the RTextTools team has refined the API and implemented a number of features. Some of these features include n-gram analysis, text labels, comprehensive analytics, and a streamlined interface. Our plan for the year ahead includes a major overhaul of the nine algorithms to facilitate low-memory ensemble classification. However, this goal involves more than just the RTextTools team; it requires the R machine learning community to strive for efficient supervised learning algorithms. Many R packages do not utilize compressed sparse matrices, and therefore are limited in their applications for large-N data-sets. Therefore, we aim to promote efficient practices by package developers and write several implementations of our own to push the capabilities of statistical computing in R.
Thank you for all your feedback and support as we look forward to another productive year in 2012!
To leave a comment for the author, please follow the link and comment on their blog: RTextTools: a machine learning library for text classification - Blog.
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.