An Experimental Evaluation of Similarity-Based and Embedding-Based Link Prediction Methods on Graphs

The task of inferring missing links or predicting future ones in a graph based on its current structure is referred to as link prediction. Link prediction methods that are based on pairwise node similarity are well-established approaches in the literature and show good prediction performance in many realworld graphs though they are heuristic. On the other hand, graph embedding approaches learn lowdimensional representation of nodes in graph and are capable of capturing inherent graph features, and thus support the subsequent link prediction task in graph. This paper studies a selection of methods from both categories on several benchmark (homogeneous) graphs with different properties from various domains. Beyond the intra and inter category comparison of the performances of the methods, our aim is also to uncover interesting connections between Graph Neural Network(GNN)- based methods and heuristic ones as a means to alleviate the black-box well-known limitation.

An Experimental Evaluation of Similarity-Based and Embedding-Based Link Prediction Methods on Graphs

The task of inferring missing links or predicting future ones in a graph based on its current structure is referred to as link prediction. Link prediction methods that are based on pairwise node similarity are well-established approaches in the literature and show good prediction performance in many realworld graphs though they are heuristic. On the other hand, graph embedding approaches learn lowdimensional representation of nodes in graph and are capable of capturing inherent graph features, and thus support the subsequent link prediction task in graph. This paper studies a selection of methods from both categories on several benchmark (homogeneous) graphs with different properties from various domains. Beyond the intra and inter category comparison of the performances of the methods, our aim is also to uncover interesting connections between Graph Neural Network(GNN)- based methods and heuristic ones as a means to alleviate the black-box well-known limitation.

Appraisal Study of Similarity-Based and Embedding-Based Link Prediction Methods on Graphs

The task of inferring missing links or predicting future ones in a graph based on its current structure is referred to as link prediction. Link prediction methods that are based on pairwise node similarity are well-established approaches in the literature and show good prediction performance in many real-world graphs though they are heuristic. On the other hand, graph embedding approaches learn low-dimensional representation of nodes in graph and are capable of capturing inherent graph features, and thus support the subsequent link prediction task in graph. This appraisal paper studies a selection of methods from both categories on several benchmark (homogeneous) graphs with different properties from various domains. Beyond the intra and inter category comparison of the performances of the methods our aim is also to uncover interesting connections between Graph Neural Network(GNN)-based methods and heuristic ones as a means to alleviate the black-box well-known limitation.

Applying a Probabilistic Network Method to Solve Business-Related Few-Shot Classification Problems

It can be challenging to learn algorithms due to the research of business-related few-shot classification problems. Therefore, in this paper, we evaluate the classification of few-shot learning in the commercial field. To accurately identify the categories of few-shot learning problems, we proposed a probabilistic network (PN) method based on few-shot and one-shot learning problems. The enhancement of the original data was followed by the subsequent development of the PN method based on feature extraction, category comparison, and loss function analysis. The effectiveness of the method was validated using two examples (absenteeism at work and Las Vegas Strip hotels). Experimental results demonstrate the ability of the PN method to effectively identify the categories of commercial few-shot learning problems. Therefore, the proposed method can be applied to business-related few-shot classification problems.

Research on the tourism destination image and sustainable development of off-season ecotourism based on tourists’ perception -- A case study of Daoxiang Lake in Beijing

In this paper, we study the question of relationship and influence between tourists’ perception and the image of off-season ecotourism destination and sustainable development of tourism, which is based on the web text. Firstly, we analyze the tendency of high-frequency words as tourism perception and customer attitude using the data of network text which is from review websites. Secondly, we study the image perception differences of ecotourism destinations in low and peak seasons through text replacement, perceived category comparison, perceived emotion comparison, and social network and network semantics comparison. Finally, we provide relevant suggestions and opinions on the sustainable development of off-season ecotourism, which is from the aspects of tourism attraction management, tourism publicity and ecological environment protection.

Association of immune-related adverse events (irAEs) with improved clinical outcome in sarcoma patients treated with immune checkpoint blockade (ICB).

11510 Background: IrAEs are associated with improved clinical outcomes after treatment with ICB in select epithelial malignancies. We hypothesized that sarcoma patients (pts) treated with ICB who developed an irAE would have improved outcomes compared to pts who had no irAE. Methods: Adverse events (AEs) from 3 sarcoma-specific ICB trials (nivolumab plus NKTR-214, pembrolizumab plus epacadostat, and pembrolizumab plus T-VEC) were reviewed. AEs probably or definitely related to ICB were classified as immune- or non-immune-related by the principal investigator. Endpoints of interest included best overall response (BOR) by RECIST 1.1 (complete response [CR]/partial response [PR]), durable clinical benefit (DCB; CR/PR/stable disease [SD] ≥ 16 weeks), and progression-free survival (PFS). Outcomes were stratified by the presence or absence of ≥ 1 irAE of any grade and by grade 1-2, grade 3-4, or no irAE (three-category comparison). Results: A total of 124 pts received ICB on these studies. Median pt age was 56 (range: 13-90); 53% were male; all but one pt had a performance status of ≤ 1. BOR was PR in 12 pts, SD in 41, and PD in 69. 2 pts were not evaluable. 40 pts (32%) had ≥ 1 irAE of any grade, 6 of whom had a grade 3-4 irAE. The most common irAEs (≥ 5% of pts) were rash (15%), arthralgia (11%), myalgia (9%), pruritis (8%), and hypothyroidism (6%). The proportion of pts with a CR/PR was higher in pts with than without an irAE (18% vs. 6%, respectively; P = 0.058). A significantly higher proportion of pts with an irAE had DCB compared to those without (53% and 29%, respectively; P = 0.017). The median PFS of pts with an irAE was 16.6 months compared to 10.6 in those without (P = 0.013). The proportion of pts with a grade 3-4 irAE and a CR/PR was highest (33%) compared to pts with grade 1-2 (15%) or no irAE (6%) (P = 0.048). More pts with grade 3-4 irAE achieved DCB (67%) than grade 1-2 (50%) or no irAE (29%) (P = 0.027). Median PFS was 22.6, 15, and 10.6 weeks in the grade 3-4, grade 1-2, and no irAE groups, respectively (P = 0.047). Conclusions: Approximately one-third of advanced sarcoma pts with ICB-based immunotherapy developed an irAE. As reported previously in select carcinomas, sarcoma pts with irAEs were more likely to have clinical benefit than those without irAEs. Further research is needed to understand the mechanism behind this association and to validate these findings prospectively.

Comparisons in works of fiction of Turkish writers of the XX century

The paper deals with comparison constructions of the Turkish language on the base of works of fiction of Turkish writers of the 20th century. Reflecting figurative (simile) and non-figurative (correlation between qualitative and quantitative characteristics of two objects) collation of two or more objects comparisons in the Turkish language are examined from the lexico-semantic point of view and though their syntactic role in the sentence. Comparison remains a relevant and interesting topic of research despite the large number of works. It is a mean, a way of investigation and knowing of reality when unknown object is collated with the well-known one. Comparison may be represented as an act and a result at once. As the act of knowing reality comparison, represents such stages of exploration of the real world as identification of an object, its correlation with already known categories and groups of objects, definition of common features and juxtaposition of their intensity for further classification of the object explored. While resembling the result of act of cognition comparison demonstrates the degree of manifestation of a particular feature on a scale from complete identity to absolute difference. As an act of logical operation comparison consists of four elements, having different nominations in research works, they resemble object that is compared, object to which something is compared with, ground of comparison (some common feature for both objects) and result of comparison that describes identity, resemblance or distinctions between objects. Since logical operations of world’s investigation find reflection in language as a linguistic category comparison also has four elements in its structure. According to the lexico-semantic group of object and subject of comparison, Turkish comparisons may be divided on such groups as human-human, human-animal, nature-human. Comparisons are mostly used to express nature, human’s appearance or character, size of items, to make some estimation, to describe the way something is happening like or to resemble someone’s condition. Turkish works of fiction use both figurative comparison (simile) and non-figurative comparison. The syntactic role of comparison in a sentence depends on which member of the sentence the comparative part of the comparative construction belongs to. Mostly, comparisons act as attribute or adverbial modifier of manner, but also may act as predicate or grammatical object.

I own, therefore I am? The role of material possession for consumers’ selfperception and attitude towards non-ownership consumption modes

In reference to Belk’s theory of the extended self this thesis investigates how and to what extent the link between an object and one’s sense of self can affect consumption mode preferences. Two hypotheses are proposed: First, the level of self-extension into a material object correlates negatively with consumers’ willingness to participate in consumption modes that value access over ownership. And second, feelings of non-generosity and possessiveness have a mediating function in this relationship. Based on the premise that ownership-substituting consumption modes promise a more sustainable lifestyle the aim of this thesis is to contribute to a better knowledge on the effect of the symbolic meaning of object on and consumer preference.To empirically test the proposed theoretical model, an online survey with 130 German participants was conducted. In a correlative research design respondents were asked to evaluate a total of eight different objects in terms of personal attachment, feelings of possessiveness and non-generosity, and to state their willingness to access the objects through ownership-substituting consumption modes. Additionally, social trust and materialism were measured on a general level.The findings support the claim that the extent to which consumers perceive an object as integral to their extended self has an effect on their willingness to forsake ownership and rather access this object through renting- or sharing-offerings. Evidence is found that this effect can be mediated by possession-linked feelings such as possessiveness and non-generosity. A general model to answer the proposed hypothesis is not supported. Rather the findings raise awareness on the object-specific nature and complexity of these effects. It becomes clear that a realistic assessment of the symbolic meaning of ownership and possession requires further experimental research and multi-category comparison.

Identifying Requirements for Big Data Analytics and Mapping to Hadoop Tools

Big data is being generating in a wide variety of formats at an exponential rate. Big data analytics deals with processing and analyzing voluminous data to provide useful insight for guided decision making. The traditional data storage and management tools are not well-equipped to handle big data and its application. Apache Hadoop is a popular open-source platform that supports storage and processing of extremely large datasets. For the purposes of big data analytics, Hadoop ecosystem provides a variety of tools. However, there is a need to select a tool that is best suited for a specific requirement of big data analytics. The tools have their own advantages and drawbacks over each other. Some of them have overlapping business use cases however they differ in critical functional areas. So, there is a need to consider the trade-offs between usability and suitability while selecting a tool from Hadoop ecosystem. This paper identifies the requirements for Big Data Analytics (BDA) and maps tools of the Hadoop framework that are best suited for them. For this, we have categorized Hadoop tools according to their functionality and usage. Different Hadoop tools are discussed from the users’ perspective along with their pros and cons, if any. Also, for each identified category, comparison of Hadoop tools based on important parameters is presented. The tools have been thoroughly studied and analyzed based on their suitability for the different requirements of big data analytics. A mapping of big data analytics requirements to the Hadoop tools has been established for use by the data analysts and predictive modelers.

Retrospective of Presentations and Analysis of Experience in the Implementation of Pedagogical Support for Children's Self-Development in Supplementary Education

The article is devoted to the study of the essence of pedagogical support for self-development of an individual as an independent scientific category, comparison of its morphology with other methods of interaction between teachers and pupils that are close by nature (upbringing, pedagogical support). Attention is focused on the fact that the history of pedagogical thought and educational practice, using the term "pedagogical support of children's self-development," mostly resorts to demonstrating the accumulated pedagogical experience of implementing various educational approaches in conditions of additional education, extra-curricular activity, children's camps and profile shifts, that gives grounds to conclude that additional education is an environment that is optimal for the implementation of the ideas of pedagogical training of the self-development of the maturing personality. The results of empirical study of the problem are presented.