LDA Analysis Research Articles

토픽모델링을 활용한 북미의 퀘벡문학 연구 동향 분석

In this study, topic modeling technique, which is one of big data analysis techniques using artificial intelligence, was applied to the investigation of the research trends of Quebec literature in North America. The data collection was done through the Web of science, and 421 Quebec literature-related papers published in North America over the last 20 years were collected. The data consisted of the titles, abstracts, and keywords of these papers, and LDA, an algorithm for topic modeling was used to analyze the data. According to the Word Cloud result, it was found that the genres of ‘novel’ and ‘poetry’ were the most studied. As a result of the LDA analysis, eight topics were created, and the topics were : ‘Quebec identity and immigrant litterature’, ‘Short story and essay’, ‘Translation and various cultures’, ‘Quebec novels and authors’, ‘Contemporary Quebec theatre and drama’, ‘Poetry’, ‘History of Quebec literature’, and ‘Quebec women's literature’. The results of this study are significant in that they attempted to analyze a vast amount of literature research papers by applying big data analysis techniques based on artificial intelligence, and are expected to serve as a stepping stone for similar studies in the future.

A Separation‐Sensing Platform Performing Accurate Diagnosis of Jaundice in Complex Biological Tear Fluids

AbstractThe detection of biomarkers in tears has aroused great interest owing to the advantages of non‐invasive and rapid collection. The combination of ultrasensitivity and label‐free detection of surface‐enhanced Raman spectroscopy (SERS) sensors is expected to achieve real‐time diagnosis in home medical care. However, the surface of SERS sensors is susceptible to biofouling and inactivation by biological impurities in tears, resulting in rapid degradation of sensitivity, limiting the commercialization of point‐of‐care devices. Herein, a binary nanosphere array with dual properties is constructed as a separation‐sensing platform for the diagnosis of target molecules in tears. The upper part of the structure is composed of Au nanoparticles (AuNPs) and a sputtering Au layer, which can bind the target molecules that interact with Au and provide high‐strength and high‐density SERS hotspots. The lower half is an inactive SiO2 nanosphere array with periodic large pores that allows biological impurities to penetrate the lower part and be separated from the target analyte. Furthermore, this substrate was integrated into homemade tear kits, enabling simultaneous tear collection, pre‐separation, and detection. Combined with the Raman spectra of tears and LDA analysis, we successfully identified patients with jaundice in clinics. This platform is expected to provide an opportunity for early disease screening based on biological fluids.

A Separation‐Sensing Platform Performing Accurate Diagnosis of Jaundice in Complex Biological Tear Fluids

The detection of biomarkers in tears has aroused great interest owing to the advantages of non-invasive and rapid collection. The combination of ultrasensitivity and label-free detection of surface-enhanced Raman spectroscopy (SERS) sensors is expected to achieve real-time diagnosis in home medical care. However, the surface of SERS sensors is susceptible to biofouling and inactivation by biological impurities in tears, resulting in rapid degradation of sensitivity, limiting the commercialization of point-of-care devices. Herein, a binary nanosphere array with dual properties is constructed as a separation-sensing platform for the diagnosis of target molecules in tears. The upper part of the structure is composed of Au nanoparticles (AuNPs) and a sputtering Au layer, which can bind the target molecules that interact with Au and provide high-strength and high-density SERS hotspots. The lower half is an inactive SiO2 nanosphere array with periodic large pores that allows biological impurities to penetrate the lower part and be separated from the target analyte. Furthermore, this substrate was integrated into homemade tear kits, enabling simultaneous tear collection, pre-separation, and detection. Combined with the Raman spectra of tears and LDA analysis, we successfully identified patients with jaundice in clinics. This platform is expected to provide an opportunity for early disease screening based on biological fluids.

Spatio-Temporal Patterns of Fitness Behavior in Beijing Based on Social Media Data

Fitness is an important way to ensure the health of the population, and it is important to actively understand fitness behavior. Although social media Weibo data (the Chinese Tweeter) can provide multidimensional information in terms of objectivity and generalizability, there is still more latent potential to tap. Based on Sina Weibo social media data in the year 2017, this study was conducted to explore the spatial and temporal patterns of urban residents’ different fitness behaviors and related influencing factors within the Fifth Ring Road of Beijing. FastAI, LDA, geodetector technology, and GIS spatial analysis methods were employed in this study. It was found that fitness behaviors in the study area could be categorized into four types. Residents can obtain better fitness experiences in sports venues. Different fitness types have different polycentric spatial distribution patterns. The residents’ fitness frequency shows an obvious periodic distribution (weekly and 24 h). The spatial distribution of the fitness behavior of residents is mainly affected by factors, such as catering services, education and culture, companies, and public facilities. This research could help to promote the development of urban residents’ fitness in Beijing.

Progress and Knowledge Transfer from Science to Technology in the Research Frontier of CRISPR Based on the LDA Model

Abstract Purpose This study explores the underlying research topics regarding CRISPR based on the LDA model and figures out trends in knowledge transfer from science to technology in this area over the latest 10 years. Design/methodology/approach We collected publications on CRISPR between 2011 and 2020 from the Web of Science, and traced all the patents citing them from lens.org. 15,904 articles and 18,985 patents in total are downloaded and analyzed. The LDA model was applied to identify underlying research topics in related research. In addition, some indicators were introduced to measure the knowledge transfer from research topics of scientific publications to IPC-4 classes of patents. Findings The emerging research topics on CRISPR were identified and their evolution over time displayed. Furthermore, a big picture of knowledge transition from research topics to technological classes of patents was presented. We found that for all topics on CRISPR, the average first transition year, the ratio of articles cited by patents, the NPR transition rate are respectively 1.08, 15.57%, and 1.19, extremely shorter and more intensive than those of general fields. Moreover, the transition patterns are different among research topics. Research limitations Our research is limited to publications retrieved from the Web of Science and their citing patents indexed in lens.org. A limitation inherent with LDA analysis is in the manual interpretation and labeling of “topics”. Practical implications Our study provides good references for policy-makers on allocating scientific resources and regulating financial budgets to face challenges related to the transformative technology of CRISPR. Originality/value The LDA model here is applied to topic identification in the area of transformative researches for the first time, as exemplified on CRISPR. Additionally, the dataset of all citing patents in this area helps to provide a full picture to detect the knowledge transition between S&T.

Analysis of the text of the FQP for automated standard control of documents

This publication focuses on underdevelopment the possibilities of machine learn-ing to help students prepare their final qualifying paper. Purpose of the study: present the possibilities of machine learning for processing final qualifying paper texts and checking them for compliance with the requirements. The article shows the possibilities of distributing work by topic, which can help students in finding materials on their topic and algorithms for extracting and analyzing text in Rus-sian for further analysis. The research is carried out on the basis of the CRISP DM methodology and describes in detail all the necessary research steps. The pa-per shows the process of extracting text from pdf and docx files; the necessary methods of text preprocessing for further analysis; and demonstrates the capabili-ties of machine learning algorithms using the example of LDA analysis.

Predicting chilling requirement of peach floral buds using electronic nose

Chilling requirement is an important agronomic trait of peach and other deciduous fruit trees. It is a quantitative indicator of natural dormancy and its release in a low temperature environment. It is directly related to the distribution area of fruit trees and the success of cultivation. Therefore, chilling requirement evaluation is very important for cultivar breeding and plant introduction. The existing evaluation method has obvious disadvantages such as harsh conditions, complex operation and time-consuming, which greatly prolong the evaluation process. Chilling requirement of peach floral buds could be quickly estimated by a portable E-nose (PEN3.5). The ratio R of the actual accumulated chilling hours during any period of the dormancy and release process to the chilling requirement was used to indicate the stage of chilling accumulation. 5 stages were set as the R value was 0 (StageⅠ), 0.4 (Stage Ⅱ), 0.8 (Stage Ⅲ), 1.0 (Stage Ⅳ) and 1.2 (Stage Ⅴ). Volatile compounds from the floral buds of 9 cultivars with known chilling requirements at the five stages were tested, and the signals of 10 types of sensors were different. Broad-methane gradually decreased during the dormancy process, and reached the bottom at dormancy release, with a significant difference in each stage (p<0.05), while it increased significantly after dormancy release (p<0.05). The sulphur-organic compound showed a significant decrease in the early and mid-term of dormancy (p<0.05), while stayed in relatively high content during and after the release (p<0.05). Methane-aliph slightly increased during the dormancy and reached the highest content at dormancy release, which was significantly different from that in the initial period of dormancy (p<0.05), and significantly decreased after the dormancy was released (p<0.05). So the three types of compounds could be used as the metabolic markers for different stages of dormancy and release. PCA and LDA analysis indicated that the 5 stages could be clearly distinguished. PLSR analysis was performed on the R value and volatile compounds content indicating by the response of sensors, and the quantitative model was established. Cultivars with different chilling requirements were used to validate the model, and it was found that there was no significant difference between predicted chilling accumulation and chilling requirement estimated by traditional method (p>0.05). It proved that the model established in this paper can be used to evaluate the chilling requirement of peach floral buds. This is the first time that E-nose detection and PLSR quantitative analysis model are combined to be used in the estimation of chilling requirements and have achieved efficient and convenient evaluation. It has a high potential to be applied in the evaluation of peach chilling requirements in the future. And for other deciduous fruit trees or plants, the innovative method is of great reference significance.

Analysis of Issues Based on Disaster Type Using Text Mining and a Semantic Network

Recently, the government has been increasingly communicating with the public in response to their opinions on state administration and policy projects. To examine the practicality of the public’s suggestions, this study investigated issues by disaster type, based on information from major media channels and comment data from the news. An analysis of the frequency of appearance, text mining (TF-IDF, LDA, and sentiment analysis), and the semantic network was performed by extracting the comment data of articles on the themes of “disaster” and “evacuation,” published from January 2010 to May 2020. The analysis results showed that news articles centered on these themes increased rapidly from 2017. The main disasters in Korea were those of “fire,” “typhoon,” “forest fire,” “radioactivity,” and “earthquake,” in order of enormity. Of the total negative words pertaining to “radioactivity” disasters, 43% were negative-sentiment words, and the semantic network analysis revealed that the terms “typhoon,” “forest fire,” and “earthquake” were connected to “radioactivity” disasters. This study is meaningful as it identifies issues by type of disaster and factors of anxiety expressed by the public using news and comment data, without conducting surveys and interviews.

Fermentation of blueberry and blackberry juices using Lactobacillus plantarum, Streptococcus thermophilus and Bifidobacterium bifidum: Growth of probiotics, metabolism of phenolics, antioxidant capacity in vitro and sensory evaluation

In this study, three potential probiotic strains were selected to ferment blueberry and blackberry juices. The viable cell counts of selected strains were increased by 0.4–0.7 log CFU/mL in berry juices environments after 48-h fermentation. Meanwhile, the contents of cyanindin-3-glucoside and peonidin-3-glucoside decreased over 30%. Heatmap presented an upgrade trend of syringic acid, ferulic acid, gallic acid and lactic acid during fermentation. However, the contents of p-coumaric acid, protocatechuic acid, chlorogenic acid, critic acid and malic acid showed downgrade trend. The metabolism of phenolics probably contributed to the enhancement of the ABTS radical scavenging activity (40%–60%) in fermented berry juices. Moreover, the three strains presented different capacities on changing the quality of berry juices according to the PCA and LDA analysis. The contents of individual organic acids had positive correlations with sensory quality, especially for sourness. Overall, probiotic fermentation could improve the sensory quality of berry juices.

When Early Adopters Learn From the Followers: The Cryptocurrency Return Predictability of GBTC Discount and Premium

We show that change in Grayscale Bitcoin Trust premium is the single most significant predictor of Bitcoin daily return. This sentiment measure is similar to the closed-end fund discount measure as in Baker and Wurgler (2006), but more likely to reflect the excess demand from traditional investors than from blockchain specialists. Although there is a substantial variation in Bitcoin price quotes worldwide, this Grayscale premium and discount predict Bitcoin daily return for the most liquid Bitcoin exchanges. Using K-means clustering and LDA analysis, we find that this predictability is especially significant when there is a large variation in bullish and bearish market sentiment, innovation regarding CBDC, regulations on crypto exchanges, but not when there is innovation regarding blockchain technology or bitcoin mining. A simple long and short strategy based on this signal generates a daily alpha of 40 bps. These findings suggest that Bitcoin prices react with a delay to the information contained in the sentiment of traditional investors and investors who are constrained from directly holding Bitcoin.

Profiling changes in primary metabolites and antioxidants during apple fruit decay caused by Penicillium crustosum

Penicillium crustosum is one of the causal agents of apple fruit blue mold decay. It produces mycotoxins and it is widely distributed, although a less aggressive pathogen than P. expansum. Changes in sugars (sucrose, glucose, and fructose), organic acids (malic and shikimic), and antioxidants (total glutathione, ascorbate, and phenolics) were monitored in P. crustosum infected apples at 6 and 24 hpi, 3, 5, 7, and 11 dpi. The initial stage of infection (6 hpi) was characterized by a decline in fructose, malic acid, and total phenolic contents, whereas the total glutathione considerably increased. Twenty-four hpi and 3 dpi the strong activation of ascorbate-glutathione cycle correlated significantly with total phenolics and organic acid enhancement at the expense of sugars. At 5 dpi the progress of visible symptoms caused a temporary breakdown of apple defense, leading to degradation of total phenolics, glutathione, and glucose, whereas the fructose as a crucial signal molecule increased significantly. An acclimation of the defense system was achieved at 7 dpi by accumulating fructose, sucrose and antioxidants that further resulted in stable ratio of most of the investigated metabolites at 11 dpi. LDA analysis enabled us to define the fructose, glucose, malic acid and ascorbate-glutathione cycle as crucial temporary marker with balanced cross-induction playing a complex role in apple defense against P. crustosum. In summary, apple fruit reacted on P. crustosum by activating defense mechanisms quickly after inoculation (6 hpi), and stayed at the higher level until 5 dpi when temporary breakdown occurred. Apple fruit managed to recover as metabolites were fairly in balance after 11 dpi. Phenolics were viewed as a wheel of the apple fruit defense mechanism, driven by the ascorbate–glutathione system, while glutathione and fructose are believed to have had a crutial role in signaling.

Microbiome in Healthy Women Between Two Districts With Different Air Quality Index.

Although the diversity and abundance of skin microbiome are mainly determined by intrinsic factors, including gender, age, anatomical site, and ethnicity, we question whether facial microbiome could be affected by long-term exposure to airborne pollution. Using 16S ribosomal RNA (rRNA) gene amplicon sequencing, we analyzed the facial bacterial microbiome of healthy and young Chinese women (25–35 years old) between two districts with different air quality indices (AQIs) in Zhejiang Province. The overall microbiome structure was obviously different between these two districts. It revealed an increase in both the abundance and diversity of facial bacterial microbiome in Hangzhou (HZ) with higher AQI compared with those in Yunhe (YH) with lower AQI. Linear discriminant analysis (LDA) and Lefse analysis identified a total of 45 genera showing significant overrepresentation in the HZ group. Furthermore, PICRUSt analysis showed that functional pathways associated with metabolism of saturated fatty acid were relatively more predominant in the HZ group, whereas those with DNA repair or mitochondrial DNA replication were more predominant in the YH group. Our present data can provide useful information for further researches on the composition and function of the skin microbiome related to air pollution factors as well as for the development of therapeutic agents targeting the microbes and their metabolites to resist damages of airborne pollutants.

Information Revealed through the Regulatory Process: Interactions between the SEC and Companies ahead of Their IPO

AbstractWe analyze communications between the SEC and firms prior to IPOs using LDA analysis and KL divergence. The SEC’s concerns closely map onto the regulator’s stated mandate: companies increase prospectus disclosures on precise topics of SEC concern. Revenue recognition is the dominant topic of SEC concern, and it is not independently discovered by investors. Increased SEC concern about it is associated with greater secondary sales, lower post-IPO liquidity, lower post-IPO returns, and a higher probability of withdrawal. The regulator’s role during the capital raising process results in increased transparency but contributes to delays in the going public process.

토픽모델링을 활용한 국방정책 연구 동향 분석

National defense policies should be established in accordance with changes in the national and international security environment and strategies of neighboring countries. In addition to defense technology policies are being advanced in the field of operation. This study analyzed the trend of defense policy research using the topic modeling. Using the LDA analysis, 14 topics including defense industry organization/composition, and civil/military/barracks were derived. The IDM analysis was performed to confirm the degree of similarity to the derived topic. The trends in defense policy research identified through this study are expected to be used as basic data for improving understanding of the defense technology field and preparing institutional policies in the defense field.

Text Classification Based on LDA and Semantic Analysis

The quality of text features directly affects the text classification effect, in order to get the text features which have the high contribution to the text classification in class, this paper proposes a text classification method based on LDA model and category semantic similarity method. The method selects text document topic features by the LDA model and calculates the semantic similarity between these features and categories combined with the word vector model. According to the size of similarity, the weight of the text feature is obtained, and the text classification feature selection and text classification are realized. Finally, the feasibility, validity and correctness of the algorithm are verified by experiments.

Text Classification Based on LDA and Semantic Analysis

Text Classification Based on LDA and Semantic Analysis

Natural vegetation cover analysis in Wadi Hagul, Egypt using hyperspectral remote sensing approach

Natural vegetation of Wadi Hagul in the eastern desert of Egypt suffers from serious changes due to the uncontrolled human activities. Present research is the first attempt to study the natural vegetation cover of Wadi Hagul in regard of spectroscopic characteristics. This was achieved by analyzing the relationship between the spectral signatures of wild plants and variation of the morphological and ecological characteristics using hyperspectral sensor. ASD field Spec was used to measure the reflection of 27 plant species in different localities. The optimal waveband was firstly selected then the optimal wavelength in order to discriminate the studied taxa using ANOVA-Tukey’s and LDA analysis. The vitality of the plant was estimated using numbers of vegetation indices derived from spectroscopic measurements. Multi-temporal Landsat images were processed to produce NDVI to detect the spatiotemporal change in vegetation cover. The study analyses showed that the investigated plant species of Wadi Hagul are belonging to 19 genera and 12 families. Statistical analyses indicated that Red and NIR spectral zones are the optimal zones to differentiate various plant species. The optimal wavelength of each plant species was found variable and analyses showed that there are wide spectral unique zones for Centaurea aegyptiaca and Ochradenus baccatus, which facilitate separating these species by various sensors. Indices analyses showed that most of the investigated plant species lies in the good conditions. It can be concluded the spectral signature of the wild plant species is very necessary for documentation of the natural vegetation in such Wadies.

A bibliometric analysis of topic modelling studies (2000–2017)

Topic modelling is a powerful text mining tool that has been applied in many fields such as software engineering, political and linguistic sciences. To evaluate the development of topic modelling studies, the present study reports a bibliometric analysis of SCIE, SSCI and A&amp;HCI listed articles published from 2000 and 2017. Bibliometric indices for productive authors, countries and institutions are analysed. In addition, thematic changes concerning topic modelling are also examined. Results show that China plays a leading role in this field. Topic modelling has established itself as an important technique in not only natural and formal sciences but also social sciences. LDA, social networks and text analysis are the topics with increasing popularity, while certain models (e.g. pLSA) and applications (e.g. topic detection) are declining in popularity. The findings could help researchers optimise research topic choices, seek collaboration with appropriate partners and stay up-to-date with the development of the field.

Image caption model of double LSTM with scene factors

In this paper, an image semantic understanding model combining scene factors is proposed to solve the problem that the accuracy rate of the description sentence is low in the current image semantic understanding model which is incorrect or ignores the scene recognition. This model first identifies the corresponding theme (scene information) through the text volume of the LDA analysis corpus. We get the vocabulary used in this scene. Then we use the ResNet to extract the global feature of the image, and use the Places365-CNNs to extract the feature of the deep scene. Finally, the model uses the picture scene information and the corpus scene information. In the description statement of the picture generation, it uses the words related to the picture scene in large probability and in the statement. In the process of generation, double LSTM is used to adjust the parameters to improve the accuracy of statement generation. This model is trained and tested in the Flickr8K, Flickr30K and MSCOCO image sets. The model is verified with different evaluation methods. The experimental results show that the proposed model can effectively improve the image language compared with other models. The accuracy of meaning understanding can solve these problems effectively.

Research Topic Analysis in Engineering Management Using a Latent Dirichlet Allocation Model

Traditional journal analyses of topic trends in IS journals have manually coded target articles from chosen time periods. However, some research efforts have been made to apply automatic bibliometric approaches, such as cluster analysis and probabilistic models, to find topics in academic articles in other research areas. The purpose of this study is thus to investigate research topic trends in Engineering Management from 1998 through 2017 using an LDA analysis model. By investigating topics in EM journals, we provide partial but meaningful trends in EM research topics. The trend analysis shows that there are hot topics with increasing numbers of articles, steady topics that remain constant, and cold topics with decreasing numbers of articles.