Faster R-CNN Algorithm Based Relay Protection Platen State Identification Method

Currently, the check of the relay protection pressure plate’s throw-out status is mainly carried out manually, due to the extremely large number of decompression plates, manual methods can cause detection errors due to fatigue. This paper proposes the processing of relay protection pressure plate photographs by using image processing techniques, the Faster R-CNN image recognition algorithm uses the feature of generating detection frames directly using RPN to identify the platen throwback status of the processed platen images, greatly improving the speed and accuracy of the detection frame generation. The experimental results show that, the method proposed in this paper effectively solves the problem of errors arising from manual verification checks of platen throwbacks, reduced workload for substation staff, the platen recognition rate can be over 98% correct.

Matching Document Pairs using Multi-Feature Semantic Fusion Based on Knowledge Graph

Discriminating the homology and heterogeneity of two documents in information retrieval is very important and difficult step. Existing methods mainly focus on word-based document duplicate checking or sentence pairs matching except manual verification which need a lot of human resource cost. The word-based document duplicate checking can not judge the similarity of two documents from the semantic level and the matching sentence pair methods can not effectively mine the semantic information from a long text which is frequent retrieval results. A concept-based Multi-Feature Semantic Fusion Model (MFSFM) is proposed. It employs multi-feature enhanced semantics to construct a concept map for represent the document, and employs a multi-convolution mixed residual CNN module to introduce local attention mechanism for improve the sensitivity of conceptual boundary information. To improve the feasibility of the proposed MFSFM based on concept maps, two multi-feature document data sets are set up. Each of them consists of about 500 actual scientific and technological project feasibility reports. Experimental results based on the actual datasets show that the proposed MFSFM converges quickly while expanding the latest methods of natural language matching at the accuracy rate.

Unmanned aerial vehicle, and GIS tools, to monitor the reproduction of the flamingo Phoenicopterus chilensis (Aves: Phoenicopteridae)

Introduction: High Andean flamingos also known as parihuanas, are species of recurrent presence in the high Andean areas which find this area as an important resting, feeding and in some cases breeding area The species recorded here correspond to Phoenicoparrus jamesi, Phoenicoparrus andinus and Phoenicopterus chilensis, the latter being the most abundant and common. During the censuses performed during 2018 and 2019, in the high Andean lake of Salinas, Ramsar site, located within the Reserva Nacional Salinas y Aguada Blanca in Southern Peru, atypical behaviors of these birds were recorded in a sector of the lake, observing reproductive courtship and the settlement of colonies of P. chilensis. Objective: The study aimed to confirm and evaluate reproductive events of P. chilenesis (Chilean flamingo) through the use of an unmanned aerial vehicle (UAV) and image processing tools using geographic information systems. Methodology: Monitoring was conducted during 2018 and 2019 to breeding colonies of P. chilensis, we used a UAV Phantom 4 testing different flight altitudes to avoid disturbing the birds and performed records of aerial photographs and GIS post-processing with the creation of panchromatic images for the identification and counting of individuals and eggs automated, and manual verification. Results: During 2018 were identified nests and presence of six eggs, this occurred between March and June where the event was interrupted not observing chicks or juveniles during this period, for 2019 the breeding was more successful, where a higher number of eggs were counted (40-66) and with the formation of three reproductive colonies with 4 185 adult individuals, also verified the presence of chicks and juveniles that reached a maximum of 1 491 individuals. Conclusions: We confirmed two continuous reproductive events of P. chilensis in the Salinas lake, where during 2019 was the most successful incorporating several new individuals to the initial population, likewise the methodology applied in the image processing allowed differentiating between adult individuals and eggs but did not allow differentiating juveniles, however, the images directly acquired by the UAV allow distinguishing the types of individuals to perform a manual count.

Sentiment Analysis for Fake News Detection

In recent years, we have witnessed a rise in fake news, i.e., provably false pieces of information created with the intention of deception. The dissemination of this type of news poses a serious threat to cohesion and social well-being, since it fosters political polarization and the distrust of people with respect to their leaders. The huge amount of news that is disseminated through social media makes manual verification unfeasible, which has promoted the design and implementation of automatic systems for fake news detection. The creators of fake news use various stylistic tricks to promote the success of their creations, with one of them being to excite the sentiments of the recipients. This has led to sentiment analysis, the part of text analytics in charge of determining the polarity and strength of sentiments expressed in a text, to be used in fake news detection approaches, either as a basis of the system or as a complementary element. In this article, we study the different uses of sentiment analysis in the detection of fake news, with a discussion of the most relevant elements and shortcomings, and the requirements that should be met in the near future, such as multilingualism, explainability, mitigation of biases, or treatment of multimedia elements.

Development and implementation of an LIS-based validation system for autoverification toward zero defects in the automated reporting of laboratory test results

Background Validation of the autoverification function is one of the critical steps to confirm its effectiveness before use. It is crucial to verify whether the programmed algorithm follows the expected logic and produces the expected results. This process has always relied on the assessment of human–machine consistency and is mostly a manually recorded and time-consuming activity with inherent subjectivity and arbitrariness that cannot guarantee a comprehensive, timely and continuous effectiveness evaluation of the autoverification function. To overcome these inherent limitations, we independently developed and implemented a laboratory information system (LIS)-based validation system for autoverification. Methods We developed a correctness verification and integrity validation method (hereinafter referred to as the "new method") in the form of a human–machine dialog. The system records personnel review steps and determines whether the human–machine review results are consistent. Laboratory personnel then analyze the reasons for any inconsistency according to system prompts, add to or modify rules, reverify, and finally improve the accuracy of autoverification. Results The validation system was successfully established and implemented. For a dataset consisting of 833 rules for 30 assays, 782 rules (93.87%) were successfully verified in the correctness verification phase, and 51 rules were deleted due to execution errors. In the integrity validation phase, 24 projects were easily verified, while the other 6 projects still required the additional rules or changes to the rule settings. Taking the Hepatitis B virus test as an example, from the setting of 65 rules to the automated releasing of 3000 reports, the validation time was reduced from 452 (manual verification) to 275 h (new method), a reduction in validation time of 177 h. Furthermore, 94.6% (168/182) of laboratory users believed the new method greatly reduced the workload, effectively controlled the report risk and felt satisfied. Since 2019, over 3.5 million reports have been automatically reviewed and issued without a single clinical complaint. Conclusion To the best of our knowledge, this is the first report to realize autoverification validation as a human–machine interaction. The new method effectively controls the risks of autoverification, shortens time consumption, and improves the efficiency of laboratory verification.

Human-efficient labeling of a solar flux emergence video dataset by a deep learning model

Machine learning is becoming a critical tool for interrogation of large complex data. However, labeling large datasets is time-consuming. Here we show that convolutional neural networks (CNNs), trained on crudely labeled astronomical videos, can be leveraged to improve the quality of data labeling and reduce the need for human intervention. We use videos of the solar photospheric magnetic field, crudely labeled into two classes: emergence or non-emergence of large bipolar magnetic regions (BMRs). We train the CNN using crude labeling, manually verify, correct labeling vs. CNN disagreements, and repeat this process until convergence. This results in a high-quality labeled dataset requiring the manual verification of only ~50% of all videos. Furthermore, by gradually masking the videos and looking for maximum change in CNN inference, we locate BMR emergence time without retraining the CNN. This demonstrates the versatility of CNNs for simplifying the challenging task of labeling complex dynamic events.

Analysis of YouTube video contents on Koha and DSpace, and sentiment analysis of viewers' comments

PurposeThe purpose of the study is to evaluate certain characteristics of the videos of the software Koha and DSpace posted on YouTube. Since YouTube has the potential to provide the content creator to share their knowledge and experience through their content which has become much more beneficial to the information seeker. Nowadays, people search for queries or tutorial videos on YouTube very often to earn a better understanding of the term. Sentiment analysis of the viewers' opinion of the videos is another purpose of this study.Design/methodology/approachDataset for evaluating the characteristic of the videos of Koha and DSpace was extracted by using Webometric Analyst by creating YouTube API. Once retrieval of data was completed, a manual verification was enhanced to filter out spam videos unrelated to the scope. After the confirmation of authentic relatable videos, seeking the video's id as query, the comments per video were extracted using Webometric Analyst. For opinion mining, the Parallel Dots API web service was used in Google Sheets as an addon function. The sentiment, multilingual sentiment, emotion, intention and word frequency of the viewers' opinion was examined with the help of certain default functionalities.FindingsWebometric Analyst extracted a total of 461 and 397 videos of Koha and DSpace, respectively, uploaded on the YouTube platform. The findings of the study indicate that the growth rate of videos on Koha is decreasing, while the number of videos uploaded on DSpace is gradually increased in the last 10 years. The highest number of videos posted in 1–20 min duration category with mostly high definition (HD) with standard YouTube license and prominently in the English language. The sentiment analysis of the total extracted comments on Koha and DSpace videos found to be 2043 and 862 comments, respectively, among whom “Positive” comments are mostly found and with “Happy” emotion can be highly detected with most supportive “Feedback” intention on both Koha and DSpace videos. The top word frequency signifies that the users of both the software are using the comments section of the videos on YouTube to ask and provide troubleshooting help to each other.Research limitations/implicationsThe present study has some limitations too; the dataset for the study includes only those videos whose title, description or keywords sections had the query terms “Koha” or “DSpace” there are chances that some videos would have been left out from the dataset related to these software.Originality/valueThis is the first paper to evaluate the characteristics and sentiment of both the videos Koha and DSpace. Through this, the popularity, likeness and dislike and the impact of the contents of the videos uploaded will be disclosed, and creators can make an improvement by referring this, and the seekers will adapt to the use of correct and authentic information.

Comparative Analysis of Milk Triglycerides Profile between Jaffarabadi Buffalo and Holstein Friesian Cow

Milk lipids are known for a variety of biological functions, however; little is known about compositional variation across breeds, especially for Jaffarabadi buffalo, an indigenous Indian breed. Systematic profiling of extracted milk lipids was performed by mass spectrometry across summer and winter in Holstein Friesian cow and Jaffarabadi buffalo. Extensive MS/MS spectral analysis for the identification (ID) of probable lipid species using software followed by manual verification and grading of each assigned lipid species enabled ID based on (a) parent ion, (b) head group, and (c) partial/full acyl characteristic ions for comparative profiling of triacylglycerols between the breeds. Additionally, new triacylglycerol species with short-chain fatty acids were reported by manual interpretation of MS/MS spectra and comparison with curated repositories. Collectively, 1093 triacylglycerol species belonging to 141 unique sum compositions between the replicates of both the animal groups were identified. Relative quantitation at sum composition level followed by statistical analyses revealed changes in relative abundances of triacylglycerol species due to breed, season, and interaction effect of the two. Significant changes in triacylglycerols were observed between breeds (81%) and seasons (59%). When the interaction effect is statistically significant, a higher number of triacylglycerols species in Jaffarabadi has lesser seasonal variation than Holstein Friesian.

RDP5: A computer program for analysing recombination in, and removing signals of recombination from, nucleotide sequence datasets

For the past 20 years the recombination detection program (RDP) project has focused on the development of a fast, flexible and easy to use Windows-based recombination analysis tool. Whereas previous versions of this tool have relied on considerable user-mediated verification of detected recombination events, the latest iteration, RDP5, is automated enough that it can be integrated within analysis pipelines and run without any user input. The main innovation enabling this degree of automation is the implementation of statistical tests to identify recombination signals that could be attributable to evolutionary processes other than recombination. The additional analysis time required for these tests has been offset by algorithmic improvements throughout the program such that, relative to RDP4, RDP5 will still run up to five times faster and be capable of analysing alignments containing twice as many sequences (up to 5000) that are five times longer (up to 50 million sites). For users wanting to remove signals of recombination from their datasets before using them for downstream phylogenetics-based molecular evolution analyses, RDP5 can disassemble detected recombinant sequences into their constituent parts and output a variety of different recombination-free datasets in an array of different alignment formats. For users that are interested in exploring the recombination history of their datasets, all the manual verification, data management and data visualization components of RDP5 have been extensively updated to minimize the amount of time needed by users to individually verify and refine the program’s interpretation of each of the individual recombination events that it detects.

Performance of a Mobile Single-Lead Electrocardiogram Technology for Atrial Fibrillation Screening in a Semirural African Population: Insights From “The Heart of Ethiopia: Focus on Atrial Fibrillation” (TEFF-AF) Study (Preprint)

BACKGROUND Atrial fibrillation (AF) screening using mobile single-lead electrocardiogram (ECG) devices has demonstrated variable sensitivity and specificity. However, limited data exists on the use of such devices in low-resource countries. OBJECTIVE The goal of the research was to evaluate the utility of the KardiaMobile device’s (AliveCor Inc) automated algorithm for AF screening in a semirural Ethiopian population. METHODS Analysis was performed on 30-second single-lead ECG tracings obtained using the KardiaMobile device from 1500 TEFF-AF (The Heart of Ethiopia: Focus on Atrial Fibrillation) study participants. We evaluated the performance of the KardiaMobile automated algorithm against cardiologists’ interpretations of 30-second single-lead ECG for AF screening. RESULTS A total of 1709 single-lead ECG tracings (including repeat tracing on 209 occasions) were analyzed from 1500 Ethiopians (63.53% [953/1500] male, mean age 35 [SD 13] years) who presented for AF screening. Initial successful rhythm decision (normal or possible AF) with one single-lead ECG tracing was lower with the KardiaMobile automated algorithm versus manual verification by cardiologists (1176/1500, 78.40%, vs 1455/1500, 97.00%; <i>P</i>&lt;.001). Repeat single-lead ECG tracings in 209 individuals improved overall rhythm decision, but the KardiaMobile automated algorithm remained inferior (1301/1500, 86.73%, vs 1479/1500, 98.60%; <i>P</i>&lt;.001). The key reasons underlying unsuccessful KardiaMobile automated rhythm determination include poor quality/noisy tracings (214/408, 52.45%), frequent ectopy (22/408, 5.39%), and tachycardia (&gt;100 bpm; 167/408, 40.93%). The sensitivity and specificity of rhythm decision using KardiaMobile automated algorithm were 80.27% (1168/1455) and 82.22% (37/45), respectively. CONCLUSIONS The performance of the KardiaMobile automated algorithm was suboptimal when used for AF screening. However, the KardiaMobile single-lead ECG device remains an excellent AF screening tool with appropriate clinician input and repeat tracing. CLINICALTRIAL Australian New Zealand Clinical Trials Registry ACTRN12619001107112; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=378057&amp;isReview=true