Saturday, March 25, 2017

Setting up cTakes as a VM

This will be ultra brief, as if you don't know how to do it, you can google each of these steps quite easily.

Using VirtualBox 5.1.6 r110634 (Qt5.5.1), create a new Ubuntu machine. I used Ubuntu 16.04. Name it something like cTakes RESTful Server. Add host-only adapter in addition to NAT adapter.

Install cTakes per the user manual.

Install Java 7 - Some of the software below is only compatible with Java 7.

sudo apt install maven

sudo apt-get update
sudo apt-get install mysql-server
sudo mysql_secure_installation

Install. Follow ytex setup instructions. Some of these files are getting hard to find. I have them if you can't find them - https://github.com/dirkweissenborn/ctakes-server

That's it! Now you can query cTakes as a RESTful server.

Busy

Well, I don't think anyone is reading this. That's OK. It acts as a lab notebook of sorts for me. I admit, I haven't posted much. I'm doing a lot though. I have an IRB going at WBAMC in El Paso TX. I'm using sklearn and DoD data to attempt to predict HF readmissions. The next two posts are some of the intermediate steps in that project. The first will briefly outline setting up cTakes as a VM so that you can query it from a RESTful interface. The second is a document classifier using NLP/cTakes and sklearn. I used it to classify radiology reports as echo or not as a step toward an automatic feature extraction system for the larger project. More to follow.