Textual Data Sources Research Articles

A simple tool for linking photo-identification with multimedia data to track mammal behaviour

Identifying individual animals is critical to describe demographic and behavioural patterns, and to investigate the ecological and evolutionary underpinnings of these patterns. The traditional non-invasive method of individual identification in mammals—comparison of photographed natural marks—has been improved by coupling other sampling methods, such as recording overhead video, audio and other multimedia data. However, aligning, linking and syncing these multimedia data streams are persistent challenges. Here, we provide computational tools to streamline the integration of multiple techniques to identify individual free-ranging mammals when tracking their behaviour in the wild. We developed an open-source R package for organizing multimedia data and for simplifying their processing a posteriori—“MAMMals: Managing Animal MultiMedia: Align, Link, Sync”. The package contains functions to (i) align and link the individual data from photographs to videos, audio recordings and other text data sources (e.g. GPS locations) from which metadata can be accessed; and (ii) synchronize and extract the useful multimedia (e.g. videos with audios) containing photo-identified individuals. To illustrate how these tools can facilitate linking photo-identification and video behavioural sampling in situ, we simultaneously collected photos and videos of bottlenose dolphins using off-the-shelf cameras and drones, then merged these data to track the foraging behaviour of individuals and groups. We hope our simple tools encourage future work that extend and generalize the links between multiple sampling platforms of free-ranging mammals, thereby improving the raw material needed for generating new insights in mammalian population and behavioural ecology.

Detecting geospatial location descriptions in natural language text

ABSTRACT References to geographic locations are common in text data sources including social media and web pages. They take different forms from simple place names to relative expressions that describe location through a spatial relationship to a reference object (e.g. the house beside the Waikato River). Often complex, multi-word phrases are employed (e.g. the road and railway cross at right angles; the road in line with the canal) where spatial relationships are communicated with various parts of speech including prepositions, verbs, adverbs and adjectives. We address the problem of automatically detecting relative geospatial location descriptions, which we define as those that include spatial relation terms referencing geographic objects, and distinguishing them from non-geographical descriptions of location (e.g. the book on the table). We experiment with several methods for automated classification of text expressions, using features for machine learning that include bag of words that detect distinctive words, word embeddings that encode meanings of words and manually identified language patterns that characterise geospatial expressions. Using three data sets created for this study, we find that ensemble and meta-classifier approaches, that variously combine predictions from several other classifiers with data features, provide the best F-measure of 0.90 for detecting geospatial expressions.

Deep learning for emotion analysis in Arabic tweets

Currently, expressing feelings through social media requires great consideration as an essential part of our lives; besides sharing ideas and thoughts, we share moments and good memories. Social media such as Facebook, Twitter, Weibo, and LinkedIn, are considered rich sources of opinionated text data. Both organizations and individuals are interested in using social media to analyze people's opinions and extract sentiments and emotions. Most studies on social media analysis mainly classified sentiment as positive, negative, or neutral classes. The challenge in emotion analysis arises because humans can express one or several emotions within one expression. Human beings can recognize these different emotions well; however, it is still not easy for an emotion analysis system. In most cases, the Arabic language used through social media is of a slangy or colloquial form, making it more challenging to preprocess and filter noise since most lemmatization and stemming tools are built on Modern Standard Arabic (MSA). An emotion analysis model has been implemented to categorize emotions. The model is a multiclass and multilabel classification problem. However, few studies have been adapted for this emotion classification problem in Arabic social media. Nearly the only work is the one of SemEval 2018 task1- sub-task E-c. Several machine learning approaches have been implemented in this task; a few studies were based on deep learning. Our model implemented a novel multilayer bidirectional long short term memory (BiLSTM) trained on top of pre-trained word embedding vectors. The model achieved state-of-the-art performance enhancement. This approach has been compared with other models developed in the same tasks using Support Vector Machines (SVM), random forest (RF), and fully connected neural networks. The proposed model achieved a performance improvement over the best results obtained for this task.

Maternal postpartum depression in relation to child undernutrition in low- and middle-income countries: a systematic review and meta-analysis.

Perinatal (antenatal and postpartum) depression is increasingly becoming a global public health problem. Studies have shown that the presence of depression during the perinatal period impact on a mother's ability to adequately provide nutritional care for herself and her infants and young children. This may thus negatively influence the nutritional status of these vulnerable children resulting in poor growth, such as stunting which is currently a public health problem in low- and middle-income countries (LMICs). The aim of this systematic review and meta-analysis was to determine the association between maternal postpartum depression and child growth in LMICs. Electronic databases were systematically reviewed by screening abstracts, titles, full text and additional data sources of eligible papers. Articles were eligible for inclusion if based on children aged 0 to 59months, if reporting maternal postpartum depression as an exposure, if based on observational studies evaluating the relation between maternal postpartum depression and child growth, if reporting at least one physical growth indicator such as stunting as an outcome, and if published after 2000. Article selection was based on the PRISMA guidelines for reporting systematic reviews. Twelve articles were included in the quantitative synthesis (12 for stunting, 8 for underweight and 2 for wasting). Maternal postpartum depression was observed to be associated with an increased risk of child stunting (1.87 (95% CI: 1.52, 2.30; I2 = 41.6%, p = 0.064)) and underweight (1.81 (95% Cl: 1.25, 2.62; I2 = 70.7%, p = 0.001)). These results were confirmed after influence analyses, with publication bias being negligible. Conclusion: Maternal depression is a risk factor for poor child growth in LMICs. This study contributes to the knowledge on this topic and calls on policymakers to improve on its understanding, screening and treatment to alleviate the potential negative impact on child growth. What is Known: • Maternal postpartum depression is currently a public health relevance in low- and middle-income countries (LMICs). • Maternal postpartum depression is a risk factor for poor nutritional status in children from LMICs. What is New: • Maternal postpartum depression is associated with an 80 to 90% increased risk of childhood stunting and underweight in LMICs. • The association between risk of childhood stunting and underweight with maternal postpartum depression is slightly influenced by geographical area, rural or urban location, by factors such as study design, study quality and sample size.

Understanding the dynamics of cost overrun triggers in highway infrastructure projects in Nigeria: a systems thinking modelling approach

PurposeEmpirical evidence suggests that many triggers influence poor cost performance in highway projects, whereas previous studies about the cost overrun triggers stem from a positivist standpoint supported by conventional statistical techniques, thus disregarding the sophisticated interactions and overall dynamics of the triggers. This study contends for a paradigm shift in investigating and understanding cost overrun triggers by adopting a holistic perspective through the lens of system thinking. This study aims to contend for a paradigm shift in investigating and understanding cost overrun triggers by adopting a holistic perspective through the lens of system thinking.Design/methodology/approachSemi-structured interviews with industry stakeholders in Nigeria were conducted buttressed by textual data from literature sources and project documents. Data analysis based on a developed data compatible coding framework and causal relations from textual data sources was used to develop a causal loop diagram depicting the interactions of the triggers which were validated by experts.FindingsThe analysis of the causal loop diagrams (CLDs) allowed identification of action points used to suggest changes for improved cost-effective highway project delivery. Among the suggested interventions are ensuring the provision of adequate funding prior to contract award will result in timely delivery of projects and indeed delivering key projects at the contractual agreed budget. This can be achieved through ensuring strict adherence to the provisions of Section 4 (2) (b) of the Nigerian Public Procurement Act, 2007 which stipulates that no contract should be awarded if funds are not available from the onset.Research limitations/implicationsThe study was limited to only highway infrastructure projects in Nigeria and as such caution must be taken before using the outcome of the study to other context within Nigeria and beyond.Practical implicationsFrom a practical point of view, the causal model demonstrates that this study is capable of being used to make pragmatic decisions regarding policy leverages about improving cost performance in highway projects provision in the Nigerian highway infrastructure sector of the construction industry. Moreover, it will aid a clear understanding of the key influencing triggers of cost overrun by the relevant stakeholders within the highway sector of the industry.Originality/valueThe hybrid-based approach applied in the development of CLDs in this study is expected to provide new insight into understanding the linkages, interactions, feedbacks and processes among the key cost overrun triggers and suggesting leverages for cost performance improvement within the philosophy of system thinking.

Weakly Supervised Learning for Categorization of Medical Inquiries for Customer Service Effectiveness.

With the growing unstructured data in healthcare and pharmaceutical, there has been a drastic adoption of natural language processing for generating actionable insights from text data sources. One of the key areas of our exploration is the Medical Information function within our organization. We receive a significant amount of medical information inquires in the form of unstructured text. An enterprise-level solution must deal with medical information interactions via multiple communication channels which are always nuanced with a variety of keywords and emotions that are unique to the pharmaceutical industry. There is a strong need for an effective solution to leverage the contextual knowledge of the medical information business along with digital tenants of natural language processing (NLP) and machine learning to build an automated and scalable process that generates real-time insights on conversation categories. The traditional supervised learning methods rely on a huge set of manually labeled training data and this dataset is difficult to attain due to high labeling costs. Thus, the solution is incomplete without its ability to self-learn and improve. This necessitates techniques to automatically build relevant training data using a weakly supervised approach from textual inquiries across consumers, healthcare professionals, sales, and service providers. The solution has two fundamental layers of NLP and machine learning. The first layer leverages heuristics and knowledgebase to identify the potential categories and build an annotated training data. The second layer, based on machine learning and deep learning, utilizes the training data generated using the heuristic approach for identifying categories and sub-categories associated with verbatim. Here, we present a novel approach harnessing the power of weakly supervised learning combined with multi-class classification for improved categorization of medical information inquiries.

Analysis of Human Behavior by Mining Textual Data: Current Research Topics and Analytical Techniques

The goal of this study was to conduct a literature review of current approaches and techniques for identifying, understanding, and predicting human behaviors through mining a variety of sources of textual data with a focus on enabling classification of psychological behaviors regarding emotion, cognition, and social empathy. This review was performed using keyword searches in ISI Web of Science, Engineering Village Compendex, ProQuest Dissertations, and Google Scholar. Our findings show that, despite recent advancements in predicting human behaviors based on unstructured textual data, significant developments in data analytics systems for identification, determination of interrelationships, and prediction of human cognitive, emotional and social behaviors remain lacking.

Code-switched automatic speech recognition in five South African languages

Most automatic speech recognition (ASR) systems are optimised for one specific language and their performance consequently deteriorates drastically when confronted with multilingual or code-switched speech. We describe our efforts to improve an ASR system that can process code-switched South African speech that contains English and four indigenous languages: isiZulu, isiXhosa, Sesotho and Setswana. We begin using a newly developed language-balanced corpus of code-switched speech compiled from South African soap operas, which are rich in spontaneous code-switching. The small size of the corpus makes this scenario under-resourced, and hence we explore several ways of addressing this sparsity of data. We consider augmenting the acoustic training sets with in-domain data at the expense of making it unbalanced and dominated by English. We further explore the inclusion of monolingual out-of-domain data in the constituent languages. For language modelling, we investigate the inclusion of out-of-domain text data sources and also the inclusion of synthetically-generated code-switch bigrams. In our experiments, we consider two system architectures. The first considers four bilingual speech recognisers, each allowing code-switching between English and one of the indigenous languages. The second considers a single pentalingual speech recogniser able to process switching between all five languages. We find that the additional inclusion of each acoustic and text data source leads to some improvements. While in-domain data is substantially more effective, performance gains were also achieved using out-of-domain data, which is often much easier to obtain. We also find that improvements are achieved in all five languages, even when the training set becomes unbalanced and heavily skewed in favour of English. Finally, we find the use of TDNN-F architectures for the acoustic model to consistently outperform TDNN–BLSTM models in our data-sparse scenario.

An Optimal Feature Selection with Wavelet Kernel Extreme Learning Machine for Big Data Analysis of Product Reviews

In recent times, generation of big data takes place in an exponential way from diverse textual data sources like review sites, media, blogs, etc. Sentiment analysis (SA) finds it useful to classify the opinions of the big data to different kinds ofsentiments. Therefore, SA on big data helps a business to take beneficial commercial understandings from text based content. Though several SA approaches have been presented, yet, there is a need to improve the performance of SA to interpret the customer’s feedback and increase the product quality.This paper introduces a novel social spider optimization based feature selection based wavelet kernel extreme learning machine (SSO-WKELM) model. The proposed model initially undergoes pre-processing to remove the unwanted word removal. Then, Term Frequency-Inverse Document Frequency (TF-IDF) is utilized as a feature extraction technique to extract the set of feature vectors. Besides, a social spider optimization (SSO) algorithm is utilized for feature selection process and thereby achieves improved classification performance. Subsequently, WKELM is employed as a classifier to classify the incidence of positive or negative user reviews. For experimental validation, a Product review dataset derived from Amazon along with synthetic data is used. The experimental results stated the superior classification performance of the SSO-WKELM model.

Getting Along with Relational Databases

Both relational databases (RDBs) and XML have strengths and weaknesses as data storage and modeling systems. Most researchers working with historical and literary data in the humanities would argue for the superiority of XML, since it allows unlimited nesting, linking, and complexity. Relational database proponents claim superior querying and processing speed, although recent advances in XML languages and tools have eroded that advantage. Nevertheless, RDBs remain popular and are widely used, particularly in the early stages of projects where resources and metadata are being collected, and projects may end up with both an RDB and an XML document collection. Programmers must then integrate these distinct forms of data when building project outputs. This article discusses the Digital Victorian Periodical Poetry (DVPP) project, where metadata on about 15,000 poems from nineteenth-century periodicals is captured in a MySQL database, and periodically exported to create a TEI file for each poem. Many of the poems are then transcribed and encoded. The canonical source of metadata is the RDB, while the canonical source of textual data is the TEI file. Metadata in the TEI files must be periodically updated from the RDB, without disturbing the textual encoding. Changes to the RDB data may result in changes to the id and filename of the related TEI file, so any existing TEI data is migrated to a new file, and the Subversion repository must be appropriately updated. All of this is done with XSLT and Ant.

Can Emoticons Be Used to Predict Sentiment?

Getting a machine to understand the meaning of language is a largely important goal to a wide variety of fields, from advertising to entertainment. In this work, we focus on Youtube comments from the top twohundred trending videos as a source of user text data. Previous Sentiment Analysis Models focus on using hand-labelled data or predetermined lexicon-s.Our goal is to train a model to label comment sentiment with emoticons by training on other user-generated comments containing emoticons. Naive Bayes and Recurrent Neural Network models are both investigated and im- plemented in this study, and the validation accuracies for Naive Bayes model and Recurrent Neural Network model are found to be .548 and .812.

Contextual Text Coding: A Mixed-methods Approach for Large-scale Textual Data

With the growing availability of large-scale text-based data sets, there is an increasing need for an accessible and systematic way to analyze qualitative texts. This article introduces and details the contextual text coding (CTC) method as a mixed-methods approach to large-scale qualitative data analysis. The method is particularly useful for complex text, textual data characterized by context-specific meanings and a lack of consistent terminology. CTC provides an alternative to current approaches to analyzing large textual data sets, specifically computational text analysis and hand coding, neither of which capture both the qualitative and quantitative analytical potential of large-scale textual data sets. Building on hand coding techniques and systematic sampling methods, CTC provides a clear six-step process to produce both quantitative and qualitative analyses of large-scale complex textual data sources. This article includes two examples, using projects focusing on journal and interview data, respectively, to illustrate the method’s versatility.

Artificial intelligence translation under the influence of multimedia teaching to study English learning mode

With the gradual advancement of education reform, English learning has become more and more important, and efficient and fast English learning has become a concern of people. In this study, an online translation platform based on artificial intelligence was selected for exploration. Use SQLite software database server as text data source. When the user performs English translation, the query word will be input into the translation platform, and at the same time, the artificial intelligence Google API translation technology will translate the text. Based on the collected actual user service usage records, build edge computing solutions and analyze user data input records. Use word2vec to construct the feature vector of the word, and use LSTM to construct the word ranking of the text. The word ranking method is used to predict the user's service usage, select the corresponding edge server, and combine the relevant probability model to preload the service. Combined with the edge algorithm compression data processing technology of the Internet of Things, the data is synchronized and tracked. After the data is compressed, the translated text is displayed on the application interface in the form of voice and words. The research results show that the use of edge computing of the Internet of Things increases the preservation rate of historical query records of intelligent translation by 30%, and the intelligent translation platform predicts that the matching degree of users' translation queries is 90%. The use of edge algorithm compression technology of the Internet of Things can greatly save the network traffic and bandwidth of the server. The optimized query algorithm can improve query efficiency, save query time, and increase users' enthusiasm for learning English.

Machine learning in management accounting research: Literature review and pathways for the future

This paper explores the possibilities of machine learning (ML) methods in management accounting research and showcases one future avenue in practice by applying ML-based textual literature review to ML/AI research in accounting. The review reveals that machine learning methods in management accounting (MA) are still in their infancy, and current research in accounting has progressed in and focused mainly on three areas related to ML and AI: 1) effects on the field of accounting and the development of the accounting profession, 2) textual analysis related to accounting data/reports, and 3) prediction methods. Based on our literature review and recently published related ML research from other fields (e.g., finance and economics), we suggest that the most promising areas for MA research to employ ML-methods are: exploitation of the rich potential of various textual data sources; quantifying qualitative and unstructured data (for example, text and images) to create new measures; creating better estimates and predictions; and using explainable AI to interpret ML models in detail.

Topic Modelling and Hotel Rating Prediction based on Customer Review in Indonesia

The growth of the tourism sector and the use of hotel online booking platforms lead to the creation of textual data sources in the form of customer review. Motivation of this study is to add value to the customer review, using more than 50,000 samples taken from 510 hotels across Indonesia. First added value is understanding most talked topics by hotel customers. Using topic model latent Dirichlet allocation (LDA), this study revealed that services, price/food, facility, comfort and location are the most talked topics. Secondly, numerical hotel rating is derived from textual data using ridge regression. In addition, regression coefficient indicates the sentiment of each word in the customer review. Finally, the output of this study is expected to be useful for customers in assessing hotel service quality and in making booking decisions, and for hotel operators to get additional input during management decision making.

Topic modelling and hotel rating prediction based on customer review in Indonesia

The growth of the tourism sector and the use of hotel online booking platforms lead to the creation of textual data sources in the form of customer review. Motivation of this study is to add value to the customer review, using more than 50,000 samples taken from 510 hotels across Indonesia. First added value is understanding most talked topics by hotel customers. Using topic model latent Dirichlet allocation (LDA), this study revealed that services, price/food, facility, comfort and location are the most talked topics. Secondly, numerical hotel rating is derived from textual data using ridge regression. In addition, regression coefficient indicates the sentiment of each word in the customer review. Finally, the output of this study is expected to be useful for customers in assessing hotel service quality and in making booking decisions, and for hotel operators to get additional input during management decision making.

ИСПОЛЬЗОВАНИЕ ТЕКСТ-МАЙНИНГАВ ЭКОНОМИКО-ГЕОГРАФИЧЕСКОМ ОТРАСЛЕВОМ АНАЛИЗЕ ЦЕЛЛЮЛОЗНО-БУМАЖНОЙ ПРОМЫШЛЕННОСТИ ЕВРОПЕЙСКОЙ РОССИИ

The authors show the need and some existing opportunities for analysis of non-traditional data sources to obtain a complete and more relevant picture of industries spatial development. The research methodology includes the use of text mining for economic and geographical studies. The relevance of the research is determined by insufficient completeness of official statistical data, cheapening of relevant information processing technologies and abundance of large text data sources in open access. The article discusses the role of the pulp and paper industry (as a key part of the timber industry) in economic and spatial development of modern Russia. The authors identify main trends in the economic and spatial development of the pulp and paper industry of European Russia, draw the conclusions on the expected industry trends and give recommendations for strategic management decisions to respond to industry challenges. The authors claim that the industry needs liberalization and stabilization, primarily through moratoriums on policy changes. The role of the use of big data, and in particular of text mining in economic and geographical research for reasonable and objective conclusions formation that can be used to make timely and balanced management decisions in the timber industry and the pulp and paper industry, is emphasized.

WAY TO SMART LIFE USING DATA MINING

Information offers numerous prerequisites to the end clients, for example, programming, association and stage go on. In this proposed framework, we concentrate about carefully mining the information on social media. Online media turns out to be well known from the information given by biomedical and other providers. This data is usually shared so that medical service improves, costs declines and utilises the assessment which is created by client. We suggest investigation framework that give attentions on side effects of drugs and also focus on positive and negative response. To improve health care some Clinical documents are mostly useful because these are free text data sources. Clinical reports containing data identified with manifestations and significant meds. Extracting data from enormous dataset became famous in light of the fact that clients get different thoughts from this shifted information. Data Mining and Knowledge mining became well known on the grounds that clients are aware of information and getting data of various region like wellbeing, Social, etc. After information preparing we center on clients positive and negative assessments. We exclude these feelings and discover which prescription is great; to choose this we additionally discover the symptoms of the drugs. Further we center on the manifestations of the sickness of tolerant. By taking the master specialists proposal, we rattle off the medicine of any sickness as per the side effects and we give this medicine or treatment to the client on our gathering. We can grow our examination into Data and Knowledge mining of online media and takes the client’s reviews on different medications of Illness. This day by day refreshed information serves to drug industry, specialists, emergency clinics, and clinical staff, for viable future medicines. KEYWORDS: Information mining, Complex organizations, social figuring, Data mining, semantic Web.

An Art of Web Services using Web Mining and Semantic Web Techniques

Todays web browsers serve as an easy access to numerous sources of text and multimedia data. More than a billion pages are indexed by search engines, and finding the desired information is not an easiest task. Over the last decade, there is an explosive growth in the information available on the World Wide Web (WWW). The objective of this paper is to provide an outline of web mining, its various classifications, its subtasks, and to give a perspective to the research community about the potential of applying techniques to extract meaningful patterns. This paper also gives information in the area of web services, semantic web mining and comparison of traditional web applications and semantic web applications thereby providing the guidelines for future research in the area of web services using web mining and semantic web.

Discovering latent themes in traffic fatal crash narratives using text mining analytics and network topology

The proliferation of digital textual archives in the transportation safety domain makes it imperative for the inventions of efficient ways of extracting information from the textual data sources. The present study aims at utilizing crash narratives complemented by crash metadata to discern the prevalence and co-occurrence of themes that contribute to crash incidents. Ten years (2009–2018) of Michigan traffic fatal crash narratives were used as a case study. The structural topic modeling (STM) and network topology analysis were used to generate and examine the prevalence and interaction of themes from the crash narratives that were mainly categorized into pre-crash events, crash locations and involved parties in the traffic crashes. The main advantage of the STM over the other topic modeling approaches is that it allows the researchers to discover themes from documents and estimate how the topic relates to the document metadata. Topics with the highest prevalence for the angle, head-on, rear-end, sideswipe and single motor vehicle crashes were crash at stop-sign, crossing the centerline, unable to stop, lane change maneuver and run-off-road crash, respectively. Eigenvector centrality measure in network topology showed that event-related topics were consistently central in articulating the crash occurrence. The centrality and association between topics varied across crash types. The efficacy of generated topics in classifying crashes by type was tested using a machine learning algorithm, Random Forest. The classification accuracy in the held-out sample ranged between 89.3 % for sideswipe crashes to 99.2 % for single motor vehicle crashes. High classification accuracy suggests that automation of crash typing and consistency checks can be accomplished effectively by using extracted latent themes from the crash narratives.