Topic modelling in precision medicine with its applications in personalized diabetes management

Abstract

Advances in Internet of Things (IoT) and analytic‐based systems in the past decade have found several applications in medical informatics, and have significantly facilitated healthcare decision making. Patients' data are collected through a variety of means, including IoT sensory systems, and require efficient, and accurate processing. Topic Modelling is an unsupervised machine learning algorithm for Natural Language Processing (NLP) that identifies relationships and associations within textual data. The application of Topic Modelling has been widely used on raw text data, where meaningful clusters (topics) are generated by the model. The purpose of this paper is to explore the varying methods of Topic Modelling, mostly the Latent Dirichlet allocation (LDA) model, and its applicability on personalized diabetes management. The proposed study evaluates the possibility of applying topic modelling methods on diabetes literature and genomic data in order to achieve precision medicine.

Perception of COVID-19 vaccination among Indian Twitter users: computational approach

“…Next random tweets were collected from each topic and analyzed manually to validate the topics and to understand the perception of the people in detail. If predefined training data for topic classification of documents is not available, topic modeling, an unsupervised machine learning technique is used to cluster text documents [ 56 , 57 ]. Topic modeling is a probability based clustering approach that aids in grouping large volume of text data in to predefined number of themes or topics [ 58 ].…”

Identifying Latent Topics and Trends in Premature Infant–Related Nursing Studies Using a Latent Dirichlet Allocation Method

“…Topic modeling has been widely applied to the analyses of raw text data in various parts of the continuum of patient care 1 . Topic modeling is primarily used to identify latent topics in a set of documents.…”

“…This technique is based on the idea that documents can be modeled as a mixture of latent topics, where each topic is a distribution over words 2 . In topic modeling, an unsupervised machine learning algorithm for natural language processing is used 1 . Natural language processing is the ability of a computer program to understand or interpret human language as it is spoken and written 3 .…”

“…Natural language processing is the ability of a computer program to understand or interpret human language as it is spoken and written 3 . An unsupervised machine learning algorithm automatically clusters word groups to best characterize a set of documents without defining topic tags or evaluating training data beforehand 1,4 . Thus, a machine learning algorithm for natural language processing breaks down human language to cluster similar expressions, identify patterns, and unlock semantic structures within text to extract insights and help make data-driven decisions 5 .…”

“…Topic modeling facilitates elucidation of the trends in the body of nursing knowledge derived from accumulated premature infant–related nursing studies. Through topic modeling, premature infant nursing studies are represented as meaningful topics generated by a model 1 . Based on this, topic modeling may help to prepare a new framework for future nursing research directions and nursing knowledge related to caring for premature infants.…”