Dynamic generation of concepts hierarchies for knowledge discovering in bio-medical linked data sets

Detailed comparison is one important sub-task of external plagiarism detection. Seed heuristic between two documents is often used in this task. Vector space model (VSM) and Jaccard coefficient are commonly used in plagiarism detection. VSM can produce high recall performance; Jaccard coefficient can produce high precision performance. In this paper, we propose a hybrid similarity measure model on the basis of the fitting function of the optimal dividing line between plagiarism and none-plagiarism where we integrates VSM and Jaccard coefficient into a unified one, our method make full use of the advantage of VSM and the Jaccard coefficient, and it can extract more reasonable heuristic seeds in the plagiarism detection. Our method is evaluated at PAN corpus of CLEF (Cross-Language Evaluation Forum) and compared with the methods based on VSM or Jaccard coefficient. Experimental results show our method can produce better performance.

Text Categorization is one of the most important tasks in information retrieval and data mining. This paper aims at investigating different variations of vector space models (VSMs) using KNN algorithm. we used 242 Arabic abstract documents that were used by (Hmeidi & Kanaan, 1997). The bases of our comparison are the most popular text evaluation measures; we use Recall measure, Precision measure, and F1 measure. The Experimental results against the Saudi data sets reveal that Cosine outperformed over of the Dice and Jaccard coefficients.

Relational databases (RDB) are widely used as a backend for information systems, and contain interesting structured data (schema and data). In the case of ontology learning, RDB can be used as knowledge source. Multiple approaches exist for building ontologies from RDB. They mainly use schema mapping to transform RDB components to ontologies. Most existing approaches do not deal with recursive relationships that can encapsulate good semantics. In this paper, two technics are proposed for transforming recursive relationships to OWL2 components: (1) Transitivity mechanism and (2) Concept Hierarchy. The main objective of this work is to build richer ontologies with deep taxonomies from RDB.

Concept hierarchy is a hierarchically organized collection of domain concepts. It is particularly useful in many applications such as information retrieval, document browsing and document classification. One of the important tasks in the construction of concept hierarchy is the identification of suitable terms with appropriate size of domain vocabulary. One way of achieving such a size is by using term reduction. The aim of this paper is to examine the effectiveness of the reduction approach to reduce the size of vocabulary using term selection methods. An experiment has been conducted on the Quran which is assumed to be a literary text. The experiment compares the entropy method, the transition point method and the hybrid of transition point and entropy methods with the Vector Space Model (VSM). Results indicate the effectiveness of the Transition Point method as compared to the others in reducing the size of the vocabulary but at the same time preserve those important terms that exist in the literary documents.

The Linking Open Data(LOD) cloud is a collection of linked Resource Description Framework (RDF) data with over 26 billion RDF triples. Consuming linked data is a challenging task because each data set in the LOD cloud has specific ontology schema, and familiarity with ontology schema is required in order to query various linked data sets. However, manually checking each data set is time-consuming, especially when many data sets from various domains are used. This difficulty can be overcome without user interaction by using an automatic method that integrates different ontology schema. In this paper, we propose a Mid-Ontology learning approach that can automatically construct a simple ontology, linking related ontology predicates (class or property) in different data sets. Our Mid-Ontology learning approach consists of three main phases: data collection, predicate grouping, and Mid-Ontology construction. Experimental results show that our Mid-Ontology learning approach successfully integrates diverse ontology schema, and effectively retrieves related information.

In the era of information explosion, it is difficult for people to obtain their desired information effectively. In tourism, a travel recommender system based on big travel data has been developing rapidly over the last decade. However, most work focuses on click logs, visit history, or ratings, and dynamic prediction is absent. As a result, there are significant gaps in both dataset and recommender models. To address these gaps, in the first step of this study, we constructed two human-annotated datasets for the travel conversational recommender system. We provided two linked data sets, namely, interaction sequence and dialogue data sets. The usage of the former data set was done to fully explore the static preference characteristics of users based on it, while the latter identified the dynamics changes in user preference from it. Then, we proposed and evaluated BERT-based baseline models for the travel conversational recommender system and compared them with several representative non-conversational and conversational recommender system models. Extensive experiments demonstrated the effectiveness and robustness of our approach regarding conversational recommendation tasks. Our work can extend the scope of the travel conversational recommender system and our annotated data can also facilitate related research.

Data on long-term prognostic factors in patients with cancer-related stroke are limited. The purpose of this study was to evaluate long-term outcomes and investigate prognostic factors of cancer-related stroke using a linked data set. Between January 2008 and December 2014, patients with cancer-related stroke (n=690) were included in a linked data set comprising the Health Insurance Review and Assessment Service claims database and the Clinical Research Center for Stroke registry data. Outcomes were long-term mortality, major bleeding including intracerebral hemorrhage and gastrointestinal bleeding, and recurrent ischemic stroke after ischemic stroke. Clinical factors related to outcomes after ischemic stroke, including vascular risk factors, acute stroke management, and types of cancer, were evaluated. Of the 690 patients with cancer-related stroke (mean age, 68.5 years and 61.7% male) during a median 1.7 year period, 401 (58.1%) patients died during follow-up duration. Multivariate Cox regression analysis revealed that old age (hazard ratio [HR], 1.30 per 10 years [95% CI, 1.17-1.45]), stroke severity (HR, 1.05 [95% CI, 1.03-1.06]), anticoagulation at discharge (HR, 1.67 [95% CI, 1.02-2.74]), and types of cancer, such as pancreatic (HR, 3.73 [95% CI, 2.70-5.16]), hepatobiliary (HR, 2.31 [95% CI, 1.78-3.00]), and lung cancers (HR, 1.73 [95% CI, 1.33-2.25]), were significantly associated with long-term mortality. Major bleeding occurred in 42 patients (6.1%), and recurrent ischemic stroke was observed in 89 patients (12.9%) during the follow-up period. Long-term outcomes of cancer-related stroke are poor. Treatment at discharge and cancer type are crucial factors associated with long-term mortality.

Today, with the development of technology, the decision-making capabilities of machines have also increased. With their high analytical skills, computers can easily catch points and relationships that may escape the human eye. Thanks to these capabilities, machines are also widely used in the field of health. For example, many machine learning techniques developed on cancer prediction have been successfully applied. Early detection of cancer is crucial to survival. In the early diagnosis of cancer, the rates of drug treatment, chemotherapy or radiotherapy that the person will be exposed to are significantly reduced and the patient gets through this process with the least amount of wear and tear. Gene Expression Cancer RNA-Seq Dataset was used in this study. This data set includes gene expression values of 5 cancer types (BRCA, KIRC, LUAD, LUSC, UCEC). DNA sequences in the dataset were analyzed using k-means and hierarchical clustering algorithms, which are unsupervised machine learning methods. The aim of the study is to develop a usable machine learning model for early detection of cancer at the gene level. Adjusted Rand Index (ARI), Silhouette Score, and Accuracy metrics were used to evaluate the analysis results. The rand index calculates similarity between clusters by counting the binaries assigned to clusters. The adjusted Rand Index is a randomly adjusted version of the Rand Index. The silhouette score indicates how well a data point fits within its own set among separated datasets. The accuracy metric is obtained as a percentage of correctly clustered data points divided by all predictions. Different connection methods are used in the hierarchical clustering algorithm. These are 'complete', 'ward', 'average' and 'single'. As a result of the study, the accuracy in the k-means algorithm was 0.990, the Adjusted Rand Index was 0.79, and the Silhouette Score was 0.14. Looking at the hierarchical clustering, ward performed the best of the four linkage methods, with an ARI score of 0.76 and a silhouette score of 0.13. As a result of the study, the accuracy of in the hierarchical clustering algorithm was 0.999.

Discovery of Abstract Knowledge from Non-Atomic Attribute Values in Fuzzy Relational Databases

In ontological theories, is-a hierarchy must represent the essential property of things and hence should be single-inheritance, since the essential property of things cannot exist in multiple. However, we cannot avoid multiperspective issues when we build an ontology because the user often want to understand things from their own viewpoints. Especially, in the Semantic Web, the variety of users causes the variety of viewpoints to capture target domains. In order to tackle this multi-perspective issue, we adopt an approach of dynamically generating is-a hierarchies according to the viewpoints of users from an ontology using single-inheritance. This article discusses a framework for dynamic is-a hierarchy generation with ontological consideration on is-a hierarchies generated by it. Then, the author shows its implementation as a new function of Hozo and its applications to a medical ontology for dynamically generation of is-a hierarchies of disease. Through the function, users can understand an ontology from a variety of viewpoints. As a result, it could contribute to comprehensive understanding of the ontology and its target world.

The traditional stator single-phase-to-ground fault protective schemes are hard to completely satisfy the requirement of selectivity for bus-connected Powerformers. In order to overcome the aforementioned problem, a new stator single-phase-to-ground fault protective scheme based on the hierarchical clustering algorithm is proposed in this paper. The direction and magnitude of zero-sequence current and leakage current are discussed and selected as feature vectors. Then, the data set are divided into two groups by hierarchical clustering algorithm, and cluster center of each group is calculated. Finally, the space relative distance among detected pattern and two cluster centers is calculated to discriminate the faulty Powerformer. The fundamental component and the third harmonic component of the data set are combined to realize 100% coverage of fault detection for the stator windings of Powerformer. Simulation results have shown that, under different fault conditions, the new scheme can distinguish internal faults from external faults reliably, and can detect in which machine a stator single-phase-to-ground fault is occurring even if the fault resistance is at 5 kΩ.

This paper examines to what extent the actual net tax-benefit position (redistribution outcome) affects support towards the welfare state besides the typical indicator for self-interest, income. This paper then explores how own and partner’s socio-demographic factors affect support based on self-assessed social risks while ideology and beliefs mitigate self-interested judgements.
 This paper integrates tax and benefit information from Luxembourg Income Study and attitudes data in European Social Survey using Statistical Matching. The technique integrates target variables that are distinctively observed in two sources using common variables which best explain the target variables in each data set. Using the linked data set, it adopts a multi-level model to explore institutional and individual factors. At the macro level, it investigates progressivity and sizes of taxes and transfers of nineteen European countries. Also, household arrangements including living with children and/or partner and demographics including age, gender, education, occupation, and own economic outlook are examined.
 Using a more direct indicator for self-interest, the net tax-benefit position, expected results of the analysis are in three folds. First, those who pay more taxes and receive less benefits are less supportive of the welfare state despite the same income level. Second, those who are more likely to benefit from the welfare state are more likely to be supportive of the welfare state. For example, those who live with children may be more supportive due to higher likelihood to receive family benefits. Similarly, subjective assessment to benefit from the welfare state due to higher social risks such as unemployment or sickness leads to more positive attitudes. Third, those who are politically left and have less meritocratic values are likely to be more supportive.
 Due to the data limitation, the impact of income-tax-benefit positions on welfare attitudes was not directly measured. This paper contributes to this discussion using the Statistical Matching and Multi-level modelling which explores interactions of actual net tax-benefit position, potential benefits based on social position/risks, and mitigating beliefs.

As more libraries begin publishing and using linked data, many in the library community may be curious about how they can start creating their own linked data sets. This article will describe the various stages in a linked data creation project, using North Carolina State University (NCSU) Libraries’ Organization Name Linked Data (ONLD) project (http://lib.ncsu.edu/ld/onld) as an example and sharing specific approaches that the project team took to produce and publish their data set. Best practices for project planning, data clean up, conversion, and publishing linked data are also discussed.

Linked Open Data initiatives have fostered the publication of Linked Data sets, as well as the deployment of publicly available SPARQL endpoints as client-server querying infrastructures to access these data sets. However, recent studies reveal that SPARQL endpoints may exhibit significant limitations in supporting real-world applications, and public linked data sets can suffer of quality issues, e.g., data can be incomplete or incorrect. We tackle these problems and propose a novel hybrid architecture that relies on shipping policies to improve the performance of SPARQL endpoints, and exploits human and machine query processing computation to enhance the quality of Linked Data sets. We report on initial empirical results that suggest that the proposed techniques overcome current drawbacks, and may provide a novel solution to make these promising infrastructures available for real-world applications.

In order to improve the recall of search results and calculate file relevancy in P2P sharing systems, a file clustering algorithm using self-organized k-means method was proposed, which is based on the hierarchical structure of file sharing directories and file names' implication of classification. With uploaded file path information built into the indexes, a tree-structure like vector space model was designed. After analyzing the advantages and shortcomings of the traditional k-means method, we implemented a revised self-organized k-means algorithm. This algorithm can easily calculate the distances among file categories and file relevancies in a same category by adjusting two thresholds called as ¿combined factor¿ and ¿correlation factor¿. From the experiment and evaluation results, this model indicated that more target files can be found and improved recall rate to 83.54% and precision of the information retrieval to 85%.