A hidden Markov model-based text classification of medical documents

Abstract

The purpose of the study is to test the application of the hidden Markov model (HMM) using prior knowledge in medical text classification (TC). HMM has been applied to a wide range of applications in information processing, but not so much in TC applications. The Medical Subject Heading (MeSH) is utilized for prior knowledge in the model. A prototype for an HMM-based TC model is designed, and an experimental model based on the prototype is implemented so as to categorize medical documents into MeSH. A subset of OHSUMED is used for the experiments. Our results show that the performance of our model is comparable to those reported in the literature.

A Hybrid Approach of Association Rule & Hidden Makov Model to Improve Efficiency Medical Text Classification

“…HMM was used as a statistical model for sequential process application in temporal pattern recognition , i.e . speech [35], handwriting [36] and bioinformatics [41], [47]. the model has been extended to the text-related task such as information retrieval [31] information extraction [13], [26] text summarization [15] text categorization [6], [12], [42], [2] also the model has been turned to the hybrid and novel model [41], [22], [21] .In [31], the research use HMM in an information retrieval model.…”

“…In the anther research [47], the research use the previous idea in a similar approach. They describe the text classification as the process of finding a relevant category c for a given documented.…”

Statistical Markovian Data Modeling for Natural Language Processing

“…Reference [41] combined text classification and HMM techniques for structuring randomized clinical trial abstracts. Reference [42] employed HMM for medical text classification. Reference [43] proposes text (sequences of pages) categorization architecture based on HMM.…”

Sentiment Classification with Supervised Sequence Embedding

“…In the rest of this manuscript, we refer to the model with uniform weights q j = 1 N (i.e., combining function is mean(·)) as SSE, while SSE-W is used to denote the model with spatial re-weighting defined in (6). In spatial re-weighting in SSE-W model, it attempts to capture longer "trends" within each document.…”

“…Notable varieties of TC include single-label, multi-label [2] or taxonomic hierarchy of labels [3]. Both generative approaches [4][5][6] and discriminative supervised methods have been applied to TC, and a few semi-supervised attempts [7] as well. Among discriminative models, support vector machines (SVM) are known for their superior performance in TC, and SC task [8,9] in particular.…”