Latent Semantic Analysis Method Research Articles

TOPIC MODELING IN COVID-19 VACCINATION REFUSAL CASES USING LATENT DIRICHLET ALLOCATION AND LATENT SEMANTIC ANALYSIS

COVID -19 vaccination is a program provided by the Indonesian government to minimize the spread of the virus. The COVID-19 vaccination program in Indonesia goes hand in hand with issues that are circulating, causing controversy and rejection of vaccination on social media, especially Twitter. There are many factors that influence vaccine rejection on Twitter, to summarize frequently discussed topics and find out hidden topics, this study uses the Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) methods from 1797 Twitter scrapping data. Both models require a set of words that have been converted into a matrix, so before conducting LDA topic modeling, the dataset will undergo a bag of word (BOW) calculation. Meanwhile, in LSA topic modeling, the existing dataset will undergo word weighting of frequently occurring words using Term Frequency - Inverse Document Frequency (TF-IDF). This study was conducted to find and summarize hidden information in the form of frequently discussed topics, thus understanding public opinions related to the COVID -19 vaccination refusal case. LDA and LSA methods will display topics based on the probability and mathematical calculations of word occurrences in each topic in the document. The topics that appear will be further analyzed through coherence score by applying a limit of 20 topics to display the best value. Further modeling experiments are carried out to display topics through LDA and LSA models, this study takes 6 topics with the highest coherence values including the right of individuals to choose whether to be vaccinated or not (0.484607), the Ribka Tjiptaning controversy (0.473368), rejection of the COVID-19 vaccine by groups represented by public figures (0.463631), punishment for non-compliance in the form of fines (0.324924), and halal certification (0.312521).

An Innovative Strategy to Anticipate Students’ Cheating: The Development of Automatic Essay Assessment on the “MoLearn” Learning Management System

A quick and effective learning assessment is needed to evaluate the learning process. Many tools currently offer automatic assessment for subjective and objective questions; however, there is no such free tool that provides plagiarism detection among students for subjective questions in a Learning Management System (LMS). This research aims to create an automatic essay assessment in MoLearn LMS that can check students’ answers. At the same time, the LMS can also anticipate online cheating. This research employed two methods comprising (1) System Development Life Cycle (SDLC) waterfall model and (2) Latent Semantic Analysis (LSA) method. The former was used to design and create essay assessment applications, and the latter was employed to check the essay's answer and automatically detect students’ plagiarism. The result showed that MoLearn LMS was successfully working on detecting students’ plagiarism, leading to better LMS innovation. Automatic essay assessment accelerates grading time by about 8.04 times that of manual grading, which takes 5.52 seconds per question.

Symbolic semantic design of industrial products based on Big data technology

Abstract Exploring the symbolic semantic design path of industrial products is to make industrial products more compatible with the diverse emotional needs of consumers. In this paper, starting from the sentiment analysis model, the PLSA-FSVM sentiment analysis method is constructed using a probabilistic latent potential semantic analysis method and support vector machine based on the Fisher kernel. The method’s validity is verified for comparative experiments and sentiment word frequency analysis evaluation. From the comparison experiments, the ten-fold cross-average precision and recall of PLSA-FSVM were 89.18% and 88.35%, respectively, 4.15% and 2.59% higher than PLSA-SVM. From the sentiment word frequency analysis, the percentages of sentiment words such as atmosphere, practical, and worthy are 23.08%, 22.59%, and 24.72%, respectively. This shows that the PLSA-FSVM sentiment analysis method can effectively realize the sentiment analysis of industrial product evaluation, promote the symbolic semantic design to be more in line with consumers’ emotional needs, and then realize the symbolic design of industrial products to reach the meaning with shape and enjoy with meaning.

Text Summarization on Verdicts of Industrial Relations Disputes Using the Cross-Latent Semantic Analysis and Long Short-Term Memory

The information presented in the documents regarding industrial relations disputes constitutes four legal disputes. However, too much information leads to difficulty for readers to find essential points highlighted in industrial relations dispute documents. This research aims to summarize automated documents of court decisions over industrial relations disputes with permanent legal force. This research involved 35 documents of court decisions obtained from Indonesia’s official Supreme Court website and employed an extractive summarization approach to summarize the documents by utilizing Cross Latent Semantic Analysis (CLSA) and Long Short-Term Memory (LSTM) methods. The two methods are compared to obtain the best results CLSA was employed to analyze the connection between phrases, requiring the ordering of related words before they were converted into a complete summary. Then, the use of LSTM is combined with the Attention module to decoder and encoder the information entered so that it becomes a form that can be understood by the system and provides a variety of splitting of documents to be trained and tested to see the highest performance that the system can generate. The research has found out that the CLSA method gave a precision of 79.1%, recall score of 39.7%, and ROUGE-1 score of 50.9%, and the use of LSTM was able to improve the performance of the CLSA method with the results obtained 93.6%, recall score of 94.5 %, and ROUGE-1 score of 93.9% on the variation of splitting 95% training and 5% testing.

A latent semantic analysis method for ranking the results of human disease search engine

The human disease search engine based on the search query about disease factors (symptom, cause, position happening symptoms, i.e.,) helps users to conveniently diagnose the disease they may have anytime, anywhere. Therefore, the disease results returned by the search engine need to be accurate and ranked reasonably so that users can know which disease has the highest probability for their search query. We propose a method to arrange the returned diseases based on the latent semantic analysis (LSA) technique. This method helps to rank the disease results reasonably and meaningfully because it not only exploits the matching term frequency-inverse document frequency (TF-IDF) scores between the disease factors in the query and the disease results, but it also exploits the implicit relationship between the disease factors in the search query and the disease factors in the disease results.

English Language Learning via YouTube: An NLP-Based Analysis of Users’ Comments

Online teaching and learning has been beneficial in facilitating learning of English as a foreign language (EFL). In online EFL learning, YouTube is one of the most utilized information and communication technology (ICT) tools because of its inherent features that make it a unique environment for learners and educators. Many interesting aspects of YouTube-based learning can be beneficial in supplementing conventional classroom methods, and, therefore, such aspects must be identified. Previous scholarly work aimed at improving YouTube learning environment was predominantly conducted manually by gathering learners’ impressions through interviews and questionnaires to analyze the differences between YouTube- and classroom-based EFL learning. However, such methods are tedious and time-consuming and can lead to results that are of less generalizable implications. User comments on YouTube channels are useful in identifying such aspects, as they present a wealth of information related to the quality of the content provided, challenges the targeted audience faces, and areas of potential improvement. Therefore, in our current study, YouTube API is used to collect the comments of three randomly selected and popular YouTube channels. Following a data cleaning process, people’s sentiments about EFL learning were first identified via a TextBlob method. Second, the automated latent semantic analysis (LSA) method of topic finding was used to collect global and open-ended topics of discussion on YouTube-based EFL learning. Users’ sentiments on the most popular topics of discussion are discussed in this paper. Further, based on the results, hypothetical findings on YouTube EFL learning are provided as recommendation for future content, including more variety of the content covered, introduction of the meanings and punctuation following words, the design of the course such that it addresses a multinational audience of any age, and targeted teaching of each variety of English, such as British and American. We also make suggestions for learners of English who wish to utilize online and offline learning, which include finding the course of interest first based on one’s needs which can be discussed with a tutor or any English teacher to optimize the learning experience, participating in fearless educator–learner interaction and engagement, and asking other EFL learners for their previous experiences with learning online in order for the learner to maximize benefit.

Determining the Best Answers for Balinese Language Problems using Latent Semantic Analysis

In Balinese, descriptions or essays are formed in an interrogative format using question words such as “akuda”, “apa”, “dija”, “kenken”, “kuda”, dan “nyen”. The assessment process on description questions tends to be more difficult and complex than multiple choice questions, this is because the description questions are described in sentence form. The solution to facilitate the assessment process on description questions can be done using automated essay scoring. Based on the results of previous studies, the Latent Semantic Analysis (LSA) method provides a better level of accuracy, because the LSA method uses the Singular Value Decomposition (SVD) method to obtain a new pattern of relationships between terms and reference terms. The data used in this study are five questions and their answer keys and there are five candidate answers for each question in Balinese. Based on the tests that have been carried out, the method used obtained an overall average accuracy of 70.26%, this shows that the LSA method can be used well in the assessment process or automatic essay assessment.

Implementasi Metode Latent Semantic Analysis Pada Peringkasan Artikel Bahasa Indonesia Menggunakan Pendekatan Steinberger Jezek

Currently, various types of news media that are developing have their own challenges, one of them is the increasing number of digital or online media. The increasing number of digital media in the form of online news sites not only has a positive impact on society, but also poses challenges. The more the number of online news portals increases, the more information is spread out. With a variety of topics and diverse information, it creates problems for readers to make time efficient in reading in full to obtain the desired information. This makes the phenomenon of text summarization very important. The purpose of this study is to analyze the Latent Semantic Analysis method in the text summarization process. The dataset used comes from previous research which includes 100 Indonesian news articles from various online news portals. In the experiment process, the Steinberger and Jezek techniques are used for the sentence selection process. The summary process is carried out using three types of compression rates, namely 10%, 30% and 50%. The best results obtained in this study are the precision value of 20% at a 10% compression rate, a recall value of 42% at a 50% compression rate, and an f-measure value of 22% at a 50% compression rate.

Keyword Extraction – Comparison of Latent Dirichlet Allocation and Latent Semantic Analysis

The main aim of the present study is to compare the keywords extracted from abstracts and full length text of scientific research papers. In addition to that, here, we compare Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA) to identify better performer for keyword extraction. This comparative study is divided into three levels, In the first level, scientific research articles on topics such as Indian Economic growth, GDP, Economic Slowdown etc. were collected and abstracts and full length text was extracted from the sources and pre-processed to remove the words and characters which were not useful to obtain the semantic structures or necessary patterns to make the meaningful corpus. In the second level, the pre-processed data were converted into a bag of words and numerical statistic TF-IDF (Term Frequency – Inverse Document Frequency) is used to assess how relevant a word is to a document in a corpus. In the third level, in order to study the feasibility of the Natural Language Processing (NLP) techniques, Latent Semantic analysis (LSA) and Latent Dirichlet Allocations (LDA) methods were applied over the resultant corpus.

Extractive text summarization for scientific journal articles using long short-term memory and gated recurrent units

Along with the increasing number of scientific publications, many scientific communities must read the entire text to get the essence of information from a journal article. This will be quite inconvenient if the scientific journal article is quite long and there are more than one journals. Motivated by this problem, encourages the need for a method of text summarization that can automatically, concisely, and accurately summarize a scientific article document. The purpose of this research is to create an extractive text summarization by doing feature engineering to extract the semantic information from the original text. Comparing the long short-term memory algorithm and gated recurrent units and were used to get the most relevant sentences to be served as a summary. The results showed that both algorithms yielded relatively similar accuracy results, with gated recurrent units at 98.40% and long short-term memory at 98.68%. The evaluation method with matrix recall-oriented understudy for gisting evaluation (ROUGE) is used to evaluate the summary results. The summary results produced by the LSTM model compared to the summary results using the latent semantic analysis (LSA) method were then obtained recall values at ROUGE-1, ROUGE-2, and ROUGE-L respectively were 76.25%, 59.49%, and 72.72%.

Key audit risks and audit procedures during the initial year of the COVID-19 pandemic: an analysis of audit reports 2019-2020

PurposeThe purpose of this study is to explore audit procedure disclosures related to key audit risks, during the prior year and the initial year of the COVID-19 outbreak, by reporting on matters published in over 3,000 Australian statutory audit reports during 2019 and 2020.Design/methodology/approachThis study partially uses latent semantic analysis methods to apply textual and readability analyses to external audit reports in Australia. The authors measure the tone of the audit reports using the Loughran and McDonald (2011) approach.FindingsThe authors find that 3% of audit procedures undertaken during 2020 were designed to address audit risks associated with the COVID-19 pandemic. As a percentage of total audit procedures undertaken during 2020, the authors find that smaller practitioners reported much less audit procedures related to COVID-19 audit risks than most larger audit firms. Finally, the textual analysis further found differences in the sentiment or tone of words used by different auditors in 2020, but differences in sentiment or tone were not found when 2020 was compared to the prior year 2019.Originality/valueThis study provides early evidence on whether auditors designed audit procedures to deal specifically with audit risks that arose due to the COVID-19 pandemic and on the extent and nature of those audit procedures. The study will help policymakers to better understand whether Key Audit Matters provided informational value to investors during a time of global crisis.

A Transformer-Based Model for Evaluation of Information Relevance in Online Social-Media: A Case Study of Covid-19 Media Posts.

Online social media has become a major source of information gathering for a huge section of society. As the amount of information flows in online social media is enormous but on the other hand, the fact-checking sources are limited. This shortfall of fact-checking gives birth to the problem of misinformation and disinformation in the case of the truthfulness of facts on online social media which can have serious effects on the wellbeing of society. This problem of misconception becomes more rapid and critical when some events like the recent outbreak of Covid-19 happen when there is no or very little information is available anywhere. In this scenario, the identification of the content available online which is mostly propagated from person to person and not by any governing authority is very needed at the hour. To solve this problem, the information available online should be verified properly before being conceived by any individual. We propose a scheme to classify the online social media posts (Tweets) with the help of the BERT (Bidirectional Encoder Representations from Transformers)-based model. Also, we compared the performance of the proposed approach with the other machine learning techniques and other State of the art techniques available. The proposed model not only classifies the tweets as relevant or irrelevant, but also creates a set of topics by which one can identify a text as relevant or irrelevant to his/her need just by just matching the keywords of the topic. To accomplish this task, after the classification of the tweets, we apply a possible topic modelling approach based on latent semantic analysis and latent Dirichlet allocation methods to identify which of the topics are mostly propagated as false information.

Topic Modelling in Knowledge Management Documents BPS Statistics Indonesia

Knowledge management is an important activity in improving the performance an organization. BPS Statistics Indonesia has recently implemented such a system to improve the quality and efficiency of business processes. The purposes of this research are: 1) implementing topic modelling on BPS Knowledge Management System to identify groups of document topics; 2) providing recommendations on which the best topic modelling; 3) building a web service function of topic modelling for BPS that includes data preprocessing function and topic group recommendation function. This study applies the Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA) topic modelling methods to determine the best grouping techniques for knowledge management systems in BPS Statistics Indonesia. The results show that the LDA model using Mallet is the best model with 25 topic groups and a coherence score of 0.4803. The performance result suggest that the best modelling method is the LDA. The LDA model is then successfully implemented in RESTful web service to provide services in the preprocessing function and topic recommendations on documents entered into the Knowledge Management System BPS.

Latent Semantic Based Fuzzy Kernel Support Vector Machine for Automatic Content Summarization

Recently, the bounteous amount of data/information has been available on the Internet which makes it very complicated to the customers to calculate the preferred data. Because the huge amount of data in a system is mandated to discover the most proper data from the corpus. Content summarization selects and extracts the related sentence depends upon the calculation of the score and rank of the corpus. Automatic content summarization technique translates from the higher corpus into smaller concise description. This chooses the very important level of the texts and implements the complete statistics summary. This paper proposes the novel technique that employs the latent semantic analysis (LSA) method where the LSA is derived from natural language processing. Also, it depends upon the particular threshold provided with the device. Statistical feature based model used to compact with inaccurate and ambiguity of the feature weights. Redundancy is removed with cosine similarity and it was presented an enhancement to the proposed method. Finally, fuzzy kernel support vector machine approach of machine learning technique is applied, so this novel model trains the classifier and predicts the statistics summary. This paper focuses to compare together with the another summarization dataset DUC (Document Understanding Conference) like ItemSum, Baseline, Summarizer, Recall Oriented Understudy for Gisting Evaluation (ROUGE) S and ROUGE L on DUC2007. The experiments and result section displays that our proposed model obtains an important performance improvement over the other classifier text summarizes.

Otomatisasi Peringkasan Teks Pada Dokumen Hukum Menggunakan Metode Latent Semantic Analysis

Pada masa sekarang, jumlah kriminalitas di Indonesia cukup banyak. Banyaknya jumlah kriminalitas di Indonesia akan berdampak pada jumlah dokumen hukum yang akan ditangani oleh aparat penegak hukum. Dalam pemahaman dokumen hukum, aparat penegak hukum seperti advokat, hakim, maupun jaksa harus membaca keseluruhan dokumen dimana akan memerlukan waktu yang cukup lama. Oleh karena itu dibutuhkan peringkasan agar aparat penegak hukum lebih mudah memahaminya. Sehingga salah satu solusi yang dibutuhkan yaitu dengan membuat sebuah ringkasan pada dokumen hukum dimana dokumen tersebut berbentuk PDF. Dalam hal meringkas teks, metode yang dapat digunakan yaitu algoritma Latent Semantic Analysis. Algoritma tersebut digunakan untuk menguraikan atau menganalisa makna yang masih tersembunyi dari suatu bahasa, kode atau jenis representasi lainnya, guna memperoleh informasi yang penting.
 Dari pengujian terhadap 10 dokumen yang diringkas oleh pakar, diperoleh hasil precision, recall, f-measure dan accuracy secara berurutan pada peringkasan teks otomatis dengan metode Latent Semantic Analysis untuk compression rate 75% yaitu 53%, 27%, 35% dan 71% lalu untuk compression rate 50% yaitu 54%, 56%, 55% dan 75%, serta untuk compression rate 25% yaitu 51%, 79%, 61% dan 75%. Berdasarkan hasil penelitian dan pengujian yang telah dilakukan dapat ditarik kesimpulan bahwa Metode Latent Semantic Analysis dapat digunakan untuk melakukan peringkasan dokumen hukum.

Job qualifications study for data science and big data professions

PurposeThe lack of clarity in defining data science is problematic in both academia and industry because the former has a need for clarity to establish curriculum guidelines in their work to prepare future professionals, and the latter has a need for information to establish clear job description guidelines to recruit professionals. This lack of clarity has resulted in job descriptions with significant overlap among different related professional groups. This study examines the industry view of five professions: statistical analysts (SAs), big data analytics professionals (BDAs), data scientists (DSs), data analysts (DAs) and business analytics professionals (BAs). The study compares the five fields with the unified backdrop of their common semantic dimensions and examines their recent dynamics.Design/methodology/approach1,200 job descriptions for the five Big Data professions (SA, DS, BDA, DA and BA) were pulled from the Monster website at four points in time, and a document library was created. The collected job qualification records were analyzed using the text analytic method of Latent Semantic Analysis (LSAs), which extract topics based on observed text usage patterns.FindingsThe findings indicated a good alignment between the industry view and the academic view of data science as a blend of statistical and programming skills. This industry view remained relatively stable during the 4 years of our study period.Originality/valueThis research paper builds upon a long tradition of related studies and commentaries. Rather than relying on subjective expertise, this study examined the job market and used text analytics to discern a space of skill and qualification dimensions from job announcements related to five big data professions.

Twitter Text Data from #Covid-19: Analysis of Changes in Time Using Exploratory Sentiment Analysis

The aim of this study is to analyse the characteristics of polish Covid-19-focused twitter posts identified via Covid-19 hashtags. The use of the popular service gives us an opportunity to monitor changing trends in users’ opinions, and changing focus of the most popular topics in the Covid-19-oriented tweets. As part of the analysis, an exploratory sentimental analysis was used to identify the main topics, emotional characteristics, and possible changes over time in the tweeter statements. The emotional characterization of the statements allowed to classify them as negative, neutral or positive. The analysis used wide range of classical text mining, sentiment analysis and latent semantic analysis methods. The use of the methods allowed obtaining the most accurate picture of the tweets. The obtained results indicate the importance of social media in creating opinions.

A comparative analysis of Latent Semantic analysis and Latent Dirichlet allocation topic modeling methods using Bible data

Objective: To compare the topic modeling techniques, as no free lunch theorem states that under a uniform distribution over search problems, all machine learning algorithms perform equally. Hence, here, we compare Latent Semantic Analysis (LSA) or Latent Dirichlet Allocation (LDA) to identify better performer for English bible data set which has not been studied yet. Methods: This comparative study divided into three levels: In the first level, bible data was extracted from the sources and preprocessed to remove the words and characters which were not useful to obtain the semantic structures or necessary patterns to make the meaningful corpus. In the second level, the preprocessed data were converted into a bag of words and numerical statistic TF-IDF (Term Frequency – Inverse Document Frequency) is used to assess how relevant a word is to a document in a corpus. In the third level, Latent Semantic analysis and Latent Dirichlet Allocations methods were applied over the resultant corpus to study the feasibility of the techniques. Findings: Based on our evaluation, we observed that the LDA achieves 60 to 75% superior performance when compared to LSA using document similarity within-corpus, document similarity with the unseen document. Additionally, LDA showed better coherence score (0.58018) than LSA (0.50395). Moreover, when compared to any word within-corpus, the word association showed better results with LDA. Some words have homonyms based on the context; for example, in the bible; bear has a meaning of punishment and birth. In our study, LDA word association results are almost near to human word associations when compared to LSA. Novelty: LDA was found to be the computationally efficient and interpretable method in adopting the English Bible dataset of New International Version that was not yet created. Keywords: Topic modeling; LSA; LDA; word association; document similarity;Bible data set

Improve topic modeling algorithms based on Twitter hashtags

Today with increase using social media, a lot of researchers have interested in topic extraction from Twitter. Twitter is an unstructured short text and messy that it is critical to find topics from tweets. While topic modeling algorithms such as Latent semantic analysis (LSA) and Latent Dirichlet Allocation (LDA) are originally designed to derive topics from large documents such as articles, and books. They are often less efficient when applied to short text content like Twitter. Luckily, Twitter has many features that represent the interaction between users. Tweets have rich user-generated hashtags as keywords. In this paper, we exploit the hashtags feature to improve topics learned from Twitter content without modifying the basic topic model of LSA and LDA. Users who share the same hashtag at most discuss the same topic. We compare the performance of the two methods (LSA and LDA) using the topic coherence ( with and without hashtags). The experiment result on the Twitter dataset showed that LSA has better coherence score with hashtags than that do not incorporate hashtags. In contrast, our experiments show that the LDA has a better coherence score without incorporating hashtags. Finally, LDA has a better coherence score than LSA and the best coherence result obtained from the LDA method was (0.6047) and the LSA method was (0.4744) but the number of topics in LDA was higher than LSA. Thus, LDA may cause the same tweets to discuss the same subject set into different clustering.

Deteksi Duplikasi Metadata File pada Media Penyimpanan menggunakan Metode Latent Semantic Analysis

Metadata files help user find relevant information, provides digital identification, archives and conserves stored files so that they are easily found and reused. The large number of data files on the storage media often makes the user unaware of the duplication and redundancy of the files that have an impact on the waste of storage media space, affecting the speed of a computer in the indexing process, finding or backing up data. This study employ the Latent Semantic Analysis method to detect file duplication and analyze the metadata of various file types in storage media. The findings showed that Latent Semantic Analysis method is able to detect duplicate file metadata in various types of storage media thereby further increasing the usability and speed of access of the data storage media.