Discovering Computer Science Research Topic Trends using Latent Dirichlet Allocation

Abstract

Before conducting a research project, researchers must find the trends and state of the art in their research field. However, that is not necessarily an easy job for researchers, partly due to the lack of specific tools to filter the required information by time range. This study aims to provide a solution to that problem by performing a topic modeling approach to the scraped data from Google Scholar between 2010 and 2019. We utilized Latent Dirichlet Allocation (LDA) combined with Term Frequency-Indexed Document Frequency (TF-IDF) to build topic models and employed the coherence score method to determine how many different topics there are for each year’s data. We also provided a visualization of the topic interpretation and word distribution for each topic as well as its relevance using word cloud and PyLDAvis. In the future, we expect to add more features to show the relevance and interconnections between each topic to make it even easier for researchers to use this tool in their research projects.

Unveiling Insights: A Knowledge Discovery Approach to Comparing Topic Modeling Techniques in Digital Health Research

“…Below is an array containing 30 sample data entries that represent the overall outcome of preprocessing: ['accelerating', 'academiaedu', 'analysing', 'apa', 'citation', 'cite', 'chicago', 'downloaded', 'formation', 'get', 'health', 'international', 'mla', 'patterns', 'public', 'pulmonary', 'rainfall', 'research', 'related', 'spatial', 'visual', 'hendra', 'rohman', 'science', 'paper', 'styles', 'tuberculosis', 'papers', 'world'] C. Exploratory Data Analysis (EDA) [36]: Calculating Coherence Score In this phase, the selection of the number of topics is based on the coherence score. The processing weighting of word analysis using Term Frequency-Inverse Document Frequency technique to reduce unnecessary words, vocabulary and eliminating noise [22]. which indicates the model's capacity to present data in a comprehensible manner for humans.…”

“…By adopting a KDF, we aim to not only assess the effectiveness of various topic modeling in digital health, but also to reveal nuanced insights that may have otherwise remained hidden. As we embark on this intellectual journey, our goal is to contribute to the refinement of methodologies in digital health research, offering a nuanced understanding of the strengths and limitations of different topic modeling approaches [22]. Through this KDF and comparative analysis, we aspire to empower researchers, practitioners, and stakeholders in the digital health domain with the knowledge needed to make informed decisions and drive advancements in the field.…”

An Emotional Analysis of Korean Topics based on Social Media Big Data Clustering

“…This process can generally be divided into two key stages: First, establishing a comprehensive set of emotional words is undertaken. Second, semantic proximity between emotion feature words is assessed, often employing techniques such as Similarity Calculation or Bootstrapping algorithms to derive an emotion semantic model [13,14].…”

Aspect-Based Sentiment Analysis for Hotel Review Using LDA, Semantic Similarity, and BERT

“…In line with the purpose of using the method in this study, latent dirichlet allocation (LDA) [10] is one of the most widely used unsupervised learning methods in topic modelling to help determine the topic of a text through hidden topics from a document. [11,12]. In previous hotel review research [7], LDA is used to extract aspect and opinion terms from hotel reviews for the ABSA proses process.…”