LDA-LFM

2021 ◽  
Vol 21 (2) ◽  
pp. 33-47
Author(s):  
Tatev Karen Aslanyan ◽  
Flavius Frasincar

Most of the existing recommender systems are based only on the rating data, and they ignore other sources of information that might increase the quality of recommendations, such as textual reviews, or user and item characteristics. Moreover, the majority of those systems are applicable only on small datasets (with thousands of observations) and are unable to handle large datasets (with millions of observations). We propose a recommender algorithm that combines a rating modeling technique (i.e., Latent Factor Model) with a topic modeling method based on textual reviews (i.e., Latent Dirichlet Allocation), and we extend the algorithm such that it allows adding extra user- and item-specific information to the system. We evaluate the performance of the algorithm using Amazon.com datasets with different sizes, corresponding to 23 product categories. After comparing the built model to four other models, we found that combining textual reviews with ratings leads to better recommendations. Moreover, we found that adding extra user and item features to the model increases its prediction accuracy, which is especially true for medium and large datasets.

Author(s):  
Hershey R. Alburo ◽  
Cherry Lyn C. Sta. Romana ◽  
Lsrmie S. Feliscuzo

The continuous pursuit of quality education has always been a concern of higher institutions. This can be seen in the way university teachers deliver academic services to the students in terms of professionalism, commitment, knowledge of the subject matter, teaching for independent learning, and management of learning. Students as recipients of these services are significant sources of information about their course interaction that takes place in an educational system. Utilizing Latent Dirichlet Allocation (LDA) algorithm and sentiment analysis through NRC emotion lexicons based on Plutchik Model, this study aimed to decipher students’ sentiments of the academic services and reveal commonalities contained in their qualitative responses. Results revealed five latent themes in the students’ responses as: The Disparity of Teaching Assignment to Professors Field of Expertise, Professors’ Expression of Willingness to Help Students in School-Related Matters, Desirable Traits Portrayed by a Professional Teacher, Professor’s Commitment and Dedication to Classroom Instruction, and Enhancement of Teaching Practices to Improve Quality of Academic Services. The results also suggest that majority of the students have a positive sentiments (64.42%), some of were negative (34.62%), and very few were neutral (0.95%). This study aimed to give inputs to any academic interventions undertaken by institution.


2012 ◽  
Vol 71 (2) ◽  
pp. 101-106 ◽  
Author(s):  
Raffaele Cioffi† ◽  
Anna Coluccia ◽  
Fabio Ferretti ◽  
Francesca Lorini ◽  
Aristide Saggino ◽  
...  

The present paper reexamines the psychometric properties of the Quality Perception Questionnaire (QPQ), an Italian survey instrument measuring patients’ perceptions of the quality of a recent hospital admission experience, in a sample of 4400 patients (Mage = 56.42 years; SD = 19.71 years, 48.8% females). The 14-item survey measures four factors: satisfaction with medical doctors, nursing staff, auxiliary staff, and hospital structures. First, we tested two models using a confirmatory factor analysis (structural equation modeling): a four orthogonal factor and a four oblique factor model. The SEM fit indices and the χ² difference suggested the acceptance of the second model. We then did a simulation using a bootstrap with 1000 replications. Results confirmed the four oblique factor solution. Third, we tested whether there were significant differences with respect to age or sex. The multivariate general linear model showed no significant differences in the factors with respect to sex or age.


Author(s):  
Yara Falmira Dianira

ABSTRACT An important factor for the success of a CSR program is effective communication. Communication will be effective if it has an impact. If the information is conveyed based on the needs, then the communication will be effective. This study aims to analyze the factors which are related to the effectiveness of CSR communication. This study used a census method to approach 37 participants who received CSR programs. The Data analysis used the Spearman rank correlation for the statistical tests. The results showed that there was a correlation between factors that have the strength of CSR companion communication (level of attractiveness of the companion, quality of message content, and sources of information) which have real communication at the level of understanding of the participants of the Kertajaya Creative Destination (KCD) CSR program. In addition, there is a real correlation the factors that have the strength of CSR companion communication (the level of credibility of the companion, the source information, and the level of the recipient) and having communication at the level of attitudes of participants in the Kertajaya Creative Destination (KCD) CSR program. However, there is no real correlation between CSR companion communication factors and participant actions.Keywords :communication effectiveness, CSR, elements of communication. ABSTRAK Faktor penting dari keberhasilan program CSR adalah komunikasi yang efektif. Komunikasi dikatakan efektif jika menimbulkan dampak. Bila informasi tersampaikan sesuai dengan kebutuhan, maka komunikasi yang dijalankan efektif. Penelitian ini bertujuan untuk menganalisis efektivitas komunikasi pendamping CSR. Penelitian ini menggunakan pendekatan sensus terhadap 37 orang peserta penerima program CSR. Analisis data menggunakan uji statistik korelasi rank Spearman. Hasil penelitian menunjukkan bahwa terdapat hubungan nyata antara faktor efektivitas komunikasi pendamping CSR (derajat daya tarik pendamping, kualaitas isi pesan, dan sumber informasi)  dengan efektivitas komunikasi pada tingkat pemahaman peserta program CSR Kertajaya Creative Destination (KCD). Selain itu, terdapat hubungan nyata antara faktor efektivitas komunikasi pendamping CSR (tingkat kredibilitas pendamping, sumber informasi, dan tingkat penerima) dengan efektivitas komunikasi pada tingkat sikap peserta program CSR Kertajaya Creative Destination (KCD). Namun, tidak terdapat hubungan nyata antara faktor efektivitas komunikasi pendamping CSR dengan tindakan peserta. Kata Kunci : CSR, efektivitas komunikasi, unsur-unsur komunikasi.


2019 ◽  
Vol 8 (3) ◽  
pp. 6634-6643 ◽  

Opinion mining and sentiment analysis are valuable to extract the useful subjective information out of text documents. Predicting the customer’s opinion on amazon products has several benefits like reducing customer churn, agent monitoring, handling multiple customers, tracking overall customer satisfaction, quick escalations, and upselling opportunities. However, performing sentiment analysis is a challenging task for the researchers in order to find the users sentiments from the large datasets, because of its unstructured nature, slangs, misspells and abbreviations. To address this problem, a new proposed system is developed in this research study. Here, the proposed system comprises of four major phases; data collection, pre-processing, key word extraction, and classification. Initially, the input data were collected from the dataset: amazon customer review. After collecting the data, preprocessing was carried-out for enhancing the quality of collected data. The pre-processing phase comprises of three systems; lemmatization, review spam detection, and removal of stop-words and URLs. Then, an effective topic modelling approach Latent Dirichlet Allocation (LDA) along with modified Possibilistic Fuzzy C-Means (PFCM) was applied to extract the keywords and also helps in identifying the concerned topics. The extracted keywords were classified into three forms (positive, negative and neutral) by applying an effective machine learning classifier: Convolutional Neural Network (CNN). The experimental outcome showed that the proposed system enhanced the accuracy in sentiment analysis up to 6-20% related to the existing systems.


2020 ◽  
Vol 24 (1) ◽  
pp. 139-152 ◽  
Author(s):  
John Armbrecht

This study focuses on the perceived quality of participatory event experiences by addressing the following question: What are the important aspects of the event experience? The aim of this research is to develop and refine a scale to measure the quality of the event experience for runners at a participatory event. The objective is to combine, apply, test, and refine the existing scales to increase our understanding of the perceived quality of events among amateur running athletes. Both affective and cognitive dimensions are included in the scale. Based on seven dimensions and 36 items, a formal scale development process is adopted. The data consist of 1,923 observations collected during a participatory event with approximately 60,000 registered participants. The seven-factor model, including immersion, surprise, participation, fun, social aspects, hedonic aspects, and service quality, was gradually revised in favor of a four-factor solution: service quality, hedonic aspects, fun, and immersion. As a result, 73.1% of the variance is extracted. This study contributes to a refined scale measuring the perceived event quality of participatory events. Service quality accounts for more than half of the variance extracted. Researchers should continue to develop research on the critical experiential dimensions in an event context. Furthermore, the links between the constructs need attention. The results suggest that event organizers should evaluate their events and event portfolios based on the scale and take actions to increase the perceived quality of these events.


Trials ◽  
2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Zhuoran Kuang ◽  
◽  
Xiaoyan Li ◽  
Jianxiong Cai ◽  
Yaolong Chen ◽  
...  

Abstract Objective To assess the registration quality of traditional Chinese medicine (TCM) clinical trials for COVID-19, H1N1, and SARS. Method We searched for clinical trial registrations of TCM in the WHO International Clinical Trials Registry Platform (ICTRP) and Chinese Clinical Trial Registry (ChiCTR) on April 30, 2020. The registration quality assessment is based on the WHO Trial Registration Data Set (Version 1.3.1) and extra items for TCM information, including TCM background, theoretical origin, specific diagnosis criteria, description of intervention, and outcomes. Results A total of 136 records were examined, including 129 severe acute respiratory syndrome coronavirus 2 (COVID-19) and 7 H1N1 influenza (H1N1) patients. The deficiencies in the registration of TCM clinical trials (CTs) mainly focus on a low percentage reporting detailed information about interventions (46.6%), primary outcome(s) (37.7%), and key secondary outcome(s) (18.4%) and a lack of summary result (0%). For the TCM items, none of the clinical trial registrations reported the TCM background and rationale; only 6.6% provided the TCM diagnosis criteria or a description of the TCM intervention; and 27.9% provided TCM outcome(s). Conclusion Overall, although the number of registrations of TCM CTs increased, the registration quality was low. The registration quality of TCM CTs should be improved by more detailed reporting of interventions and outcomes, TCM-specific information, and sharing of the result data.


2021 ◽  
pp. 0887302X2199826
Author(s):  
Muzhen Li ◽  
Li Zhao

Nowadays, more fashion companies have started to adopt various sustainability practices and communicate these practices through their annual public CSR reports. In this study, we aim to provide a holistic perspective of fashion companies’ sustainable development and investigate the sustainability practices of global fashion companies. A total of 181 CSR reports from 29 fashion companies were collected. A Dictionary approach text classification method, combined with Latent Dirichlet Allocation (LDA), a computer-assisted topic modeling algorithm, was implemented to detect and summarize the themes and keywords of detailed practices disclosed in CSR reports. The findings identified 12 main sustainability practices themes based on the triple bottom line theory and the moral responsibility of corporate sustainability theory. In general, waste management and human rights are the most frequently mentioned themes. The findings also suggest that global fashion companies adopted different sustainability strategies based on their product categories and competitive advantages.


Author(s):  
Andrea Langbecker ◽  
Daniel Catalan-Matamoros

Sources of information are a key part of the news process as it guides certain topics, influencing the media agenda. The goal of this study is to examine the most frequent voices on vaccines in the Portuguese press. A total of 300 news items were analysed via content analysis using as sources two newspapers from 2012 to 2017. Of all the articles, 97.7% included a source (n = 670). The most frequent were “governmental organisations”, “professional associations” and the “media”. Less frequent sources were “university scientists”, “governmental scientific bodies”, “consumer groups”, “doctors”, “scientific companies”, “NGOs” and “scientific journals”. Most articles used only non-scientific sources (n = 156). A total of 94 articles used both categories and 43 used exclusively scientific sources. Our findings support the assertion that media can be an instrument to disseminate information on vaccines. Nevertheless, despite being present in most articles, the number of sources per article was low, therefore not presenting a diversity of opinions and there was a lack of scientific voices, thus suggesting lower quality of the information being offered to the audience.


2020 ◽  
pp. 1-17
Author(s):  
Vikas Kumar

The quality of metadata is a crucial determinant of usability/interpretability of data. This paper draws attention to the poor quality of India’s government statistics and the paucity of metadata necessary to understand the problems. The paper suggests that there has been a decline in India both in terms of the availability and quality of metadata for key government sources of information including maps, decennial population censuses and National Sample Surveys amidst growing sophistication in the understanding of metadata. The poor quality of metadata impairs cross-sectional as well as inter-temporal comparisons and policymaking apart from concealing biases and lapses of government statisticians. The paper draws on the experience of three states – erstwhile Jammu and Kashmir, Manipur and Nagaland – where government statistics have been affected by serious errors that are not well-understood due to the lack of adequate metadata.


2008 ◽  
Vol 3 (3) ◽  
pp. 57 ◽  
Author(s):  
Carol Perryman

A Review of: Brown, Cecilia M. and Ortega, Lina. “Information-Seeking Behavior of Physical Science Librarians: Does Research Inform Practice?” College & Research Libraries (2005). 66:231-47. Objective – As part of a larger study exploring the information environments of physical science librarians (Ortega & Brown), the authors’ overall objective for this study is to profile physical science librarians’ information behaviours. The authors’ two-part hypothesis was that first, peer-reviewed journals would be preferred over all other sources for research dissemination, resembling the preferences of scientists, and second, that peer-to-peer consultation would predominate for practice-oriented decisions. Design – Mixed methods: survey questionnaire followed by citation and content analysis. Setting – Five internationally disseminated professional association electronic mailing lists whose readership comprised those with interests in science librarianship: the American Library Association (ALA) Science and Technology Section; the American Society for Information Science & Technology (ASIST) Science and Technology Information Special Interest Group; the Special Library Association (SLA) Chemistry Division and its Physics-Astronomy-Mathematics Division; and the American Geological Institute Geoscience Information Society. Subjects – Seventy-two physical science librarians voluntarily responding to an online survey. Methods – A questionnaire was distributed to inquire about physical science librarians’ professional reading practices as well as their perceptions about the applicability of research to their work. Participants were asked to rank preferences among 11 resource types as sources supporting daily business, including personal communication, conference attendance, electronic mailing lists, and scholarly journals. Differences between the mean rankings of preferences were tested for significance by applying the Friedman test with p>0.0005. Journals identified most frequently were analyzed using the Institute for Scientific Information’s (ISI) Web of Science index and Ulrich’s Periodical Index to measure proportions of research and non-research citations, as well as the general topic areas covered by the journals. Next, content analysis was performed for the years 1995, 1997, and 2000 in order to characterize research methodologies used in the previously identified journals according to a previously tested schema (Buscha & Harter). Results from this portion of the study were compared with participants’ responses about journal usage. Main Results – Librarians reported using personal communication (both face-to-face and electronic mailing lists) more frequently as a means of information gathering than professional journals, Web sites, conferences, trade publications, monographs, or ‘other’ resources. Variations in responses appeared to correlate with years in the profession and in the respondents’ time in their current positions, although there are indications that the importance of all information resources to practice and research declines over time. The relative importance of resources is also shown in time spent reading journal literature, less than 5 hours per week for 86% of participants. Conclusion – For the first hypothesis, the authors found that unlike scientists, survey participants did not prefer research publications as vehicles for dissemination of their research results. For the second, librarians ranked peer-reviewed journals third in preference after personal communication and electronic mailing lists as sources of information supporting daily practice, supporting the second hypothesis that respondents would emulate the information use practices of mathematicians.


Sign in / Sign up

Export Citation Format

Share Document