Descriptions of love can be found in a wide range of literature. The meaning of love that a reader grasps from reading a literary work is mostly the result of self-understanding and is very likely different from the one that the author tried to express. Therefore, it is interesting to explore what love is from the authors' perspective to help readers have a deeper understanding of the meaning of love written by the author. The goal of this study is to build a text analysis framework to identify common words or phrases describing love in romance literature. The proposed analysis is divided into three types, namely 1) text classification and sentiment analysis, 2) key phrase extraction, and 3) topic modeling. The evaluation is performed on 10 romance books. The results of each analysis method are measured using performance metrics as well as presented using visuals like word cloud and histogram.
TopIntroduction
The ability to perceive love is one of the important capabilities of everyone in the real world, and it may directly affect a person in establishing connections with others, and it will also have a decisive impact on the feelings of themselves and others. The characters in literary works often come from real life, and the text reflects the love in real society. It may express love directly, or it may be implicit words or phrases to express love. The latter requires us to discover and understand its deeper meaning. Therefore, it is very important to start from the source, that is, to analyze what is love from the authors’ perspective, and to convey information about love to readers. In creative writing domain, a story always requires some clues to connect the past, present, and future plot. The plot development of the works and the character setting are based on the narrative thread and serve the main thread. The focus of love is one of the most important storylines because a lot of content in literary works portray characters where feelings and emotions evolved that lead to the embodiment of love. This is especially crucial in romantic literatures, where the focuses are more on describing the relationship and love between two people. Therefore, to understand love from literature, the most basic thing is to know which words describe love, which positively affect readers and authors. According to Ebaid (2018), adjectives are defined as an indispensable part of language structure, because it is the main component of describing, identifying and modifying nouns, so adjectives are widely used to express love. Nouns are equally important, in which is it usually used in endearment to create a sense of closeness between two people. Thus, one part in this research is to describe what embody love from this perspective.
In the field of natural language processing, sentiment analysis has emerged for mining public opinion, analyzing, and extracting knowledge from the subjective information published on the Internet, while machine learning (ML) algorithms are very common for analyzing sentiment. Up to now, sentiment analysis has been a hot research field in data mining and natural language processing (NLP) recently and has been used to do the process of automatically identifying whether user-generated text expresses a positive, negative, or neutral view of an object. To analyze these sentiments, various machine learning (ML) algorithms, and natural language processing-based approaches have been used in the past. Hussain & Cambria (2018) explored a novel semi-supervised learning model based on the combined use of random projection scaling as part of a vector space model, and support vector machines to perform reasoning on a knowledge base. Perikos et al. (2021) introduced Hidden Markov Models (HMMs) for sentiment analysis, in which HMMs has the ability to produce different probability to identify the most possible identification of the sentiment. They are also able to provide visualization on every probability of the decisions made, resulting the ability to show the way the sentiments change from the start to the end of the sentence.