A Text Segmentation Approach for Automated Annotation of Online Customer Reviews, Based on Topic Modeling

Abstract

Online customer review classification and analysis have been recognized as an important problem in many domains, such as business intelligence, marketing, and e-governance. To solve this problem, a variety of machine learning methods was developed in the past decade. Existing methods, however, either rely on human labeling or have high computing cost, or both. This makes them a poor fit to deal with dynamic and ever-growing collections of short but semantically noisy texts of customer reviews. In the present study, the problem of multi-topic online review clustering is addressed by generating high quality bronze-standard labeled sets for training efficient classifier models. A novel unsupervised algorithm is developed to break reviews into sequential semantically homogeneous segments. Segment data is then used to fine-tune a Latent Dirichlet Allocation (LDA) model obtained for the reviews, and to classify them along categories detected through topic modeling. After testing the segmentation algorithm on a benchmark text collection, it was successfully applied in a case study of tourism review classification. In all experiments conducted, the proposed approach produced results similar to or better than baseline methods. The paper critically discusses the main findings and paves ways for future work.

Leveraging Topic Modelling to Analyze Biomedical Research Trends from the PubMed Database Using LDA Method

“…Alternatively, topic modelling is used to find the main topic hidden from a series of words in a large and unstructured document [29]. The topic modelling method analyzes data based on the original text regarding the relationship between topics with each other and the relationship between themes that can change at any time [30]. So, this method can be developed for searching or summarizing text contained in documents.…”

Employing a Multilingual Transformer Model for Segmenting Unpunctuated Arabic Text

“…To elaborate more, existing approaches endeavor to split a given text into subtexts while tackling different levels of language complexities. These levels range from identifying morphological and lexical cohesion boundaries within word-to-small text segments [1,4,13] (e.g., words, phrases, and simple discourse 2 sentences) to higher-level syntactic and thematic structures [22,24,27] (such as paragraphs, passages, chapters, or textual documents containing many chapters). In terms of the underlying technique used, existing text segmentation approaches also vary according to their chronological development, which can be classified into three genres:…”

“…• the early discourse segmentation approaches (some approaches are dated back to the previous decade), which often depend on extracting linguistic annotation features, including treebank structure and POS tags [4,9,10,22]; • the statistical and machine learning methods, which often manipulate text similarity metrics and distances; and recently [12][13][14]24]; Phrase to a simple sentence…”

Enhancing Text Summarization with a T5 Model and Bayesian Optimization

“…The text summary is the most complicated challenge in the field of natural language preprocessing because the source text is unstructured and there are no labels. Many researchers manual annotation as Al-Laith et al [18], Krommyda et al [19], and Hananto et al [20] do manual annotation with the aim of obtaining input text data to be used in the field of natural language processing. In carrying out text summary tasks there are many techniques that can be used, one of which is using the most popular models such as the Transformer-based Text-to-Text Transfer Transformer or the T5 model which has advantages in producing good text summaries.…”