Search Mailing List Archives
[bioontology-support] Request for the VMWare Virtual Appliance
JAHOON.KOO at UCDENVER.EDU
Wed Jan 8 12:49:22 PST 2020
This is a student worker from Health Data Compass on CU Anschutz campus.
I would like to obtain the VMWare Virtual Appliance for a project that my team is currently working on.
My BioPortal username is jahoon1998.
The primary goal of our project is to figure out a NLP tool used in health data field that is appropriate for our domain's purpose. Currently, we are exploring Google's NLP tools and academic and open source NLP tools. For Google's NLP tools, we are looking into Cloud Healthcare API, Google Cloud AutoML Natural Language, Cloud Natural Language API, and Cloud Data Loss Prevention. Also, we are studying various academic and open source NLP tools used in health data, such as BioPortal, MetaMap, and CLAMP. For each tool, the notes, measures, and topics we will explore are: organization responsible for the tool, maintenance license (academic, commercial, open source), computer language, pre-trained algorithms availability, programmable/editable/extensible, difficulty to setup and use the tool, OS compatibility, standalone and/or web, time to process and complete an NLP task, number of items (term, section, negation, temporality, etc.) found, terminologies/Vocabularies used Permit term/vocabulary customization, corpora/Lexicon used, text processing, Bag of Words (BOW) and N-Grams Document, distance or similarity metric, Topic Models, Sections/Headers/Paragraphs/Sentences, Negation, Temporality (past, current and future events), Topic Modeling, Named Entity Recognition, Sentiment Analysis, Regular Expression, Acronyms and abbreviations, Measures/Quantities, HIPAA De-Identifying Elements. We will use MIMIC-III Clinical Database 1.4 to test different NLP tools, and our health data experts will evaluate the accuracy of results and compare results from different NLP tools.
I have used the BioPortal REST API to run MIMIC dataset, but I have experienced the HTTP 414 error - URI Too Long and latency issue when annotating long texts from MIMIC dataset. Also, we will eventually run the tool with actual patient's' data and transmitting private data through REST API is not acceptable. Therefore, we prefer the local installation.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the bioontology-support