Graph Few-shot Learning with Attribute Matching

2020 ◽  
Author(s):  
Ning Wang ◽  
Minnan Luo ◽  
Kaize Ding ◽  
Lingling Zhang ◽  
Jundong Li ◽  
...  
Keyword(s):  
Attribute Matching

Matching Attributes across Overlapping Heterogeneous Data Sources Using Mutual Information

2012 ◽  
pp. 417-437
Author(s):  
Huimin Zhao
Keyword(s):  
Mutual Information ◽  
Heterogeneous Data ◽  
Data Sources ◽  
Heterogeneous Data Sources ◽  
Attribute Matching ◽  
Measure Of Dependence

Identifying matching attributes across heterogeneous data sources is a critical and time-consuming step in integrating the data sources. In this paper, the author proposes a method for matching the most frequently encountered types of attributes across overlapping heterogeneous data sources. The author uses mutual information as a unified measure of dependence on various types of attributes. An example is used to demonstrate the utility of the proposed method, which is useful in developing practical attribute matching tools.

scholarly journals Semantic Attribute Matching Networks

2019 ◽  
Author(s):  
Seungryong Kim ◽  
Dongbo Min ◽  
Somi Jeong ◽  
Sunok Kim ◽  
Sangryul Jeon ◽  
...  
Keyword(s):  
Matching Networks ◽  
Attribute Matching ◽  
Semantic Attribute

Matching Attributes across Overlapping Heterogeneous Data Sources Using Mutual Information

2010 ◽  
Vol 21 (4) ◽  
pp. 91-110 ◽  
Author(s):  
Huimin Zhao
Keyword(s):  
Mutual Information ◽  
Heterogeneous Data ◽  
Data Sources ◽  
Heterogeneous Data Sources ◽  
Attribute Matching ◽  
Measure Of Dependence

Identifying matching attributes across heterogeneous data sources is a critical and time-consuming step in integrating the data sources. In this paper, the author proposes a method for matching the most frequently encountered types of attributes across overlapping heterogeneous data sources. The author uses mutual information as a unified measure of dependence on various types of attributes. An example is used to demonstrate the utility of the proposed method, which is useful in developing practical attribute matching tools.

Quantitative Calculation Methods of Modules Matching for Product Configuration

2010 ◽  
Vol 20-23 ◽  
pp. 1391-1396
Author(s):  
Jie Hui Zou ◽  
Qun Gui Du
Keyword(s):  
Graph Theory ◽  
Design Method ◽  
Product Configuration ◽  
Calculation Methods ◽  
Matching Problems ◽  
Future Studies ◽  
Quantitative Calculation ◽  
Attribute Matching ◽  
Important Design

Modularization is an important design method cope with complicated and diversified products.Firstly, idea of quantitative calculation about modules matching base on attribute matching functions is proposed. Secondly, matching problems between two modules and among multi-modules are studied deeply by using graph theory. Then, a simple example is given to show the process of quantitative calculation and prove its feasibility. Finally, conclusions and future studies are presented.

scholarly journals A Matching Method of Heterogeneous Database based on SOM and BP Neural Network

2021 ◽  
Vol 15 ◽  
pp. 383-392
Author(s):  
Yongjie Zhu ◽  
Shenzhan Feng
Keyword(s):  
Large Scale ◽  
Recall Rate ◽  
Heterogeneous Data ◽  
Heterogeneous Databases ◽  
Heterogeneous Database ◽  
Matching Model ◽  
Bp Network ◽  
Data Matching ◽  
Matching Process ◽  
Attribute Matching

In the process of data integration among heterogeneous databases, it is significantly important to analyze the identical attributes and characteristics of the databases. However, the existing main data attribute matching model has the defects of oversize matching space and low matching precision. Therefore, this paper puts forward a heterogeneous data attribute matching model on the basis of fusion of SOM and BP network through analyzing the attribute matching process of heterogeneous databases. This model firstly matches the heterogeneous data attributes in advance by SOM network to determine the centre scope of attribute data to be matched. Secondly, the accurate match will be carried out through BP network of the standard heterogeneous data various attribute center. Finally, the matching result of the relevant actual database shows that this model can effectively reduce the matching space in the case of complex pattern. As for the large-scale data matching, the matching accuracy is relatively high. The average precision is 89.52%, and the average recall rate is 100%.

scholarly journals An Attribute-Oriented Task Scheduling Strategy for Improvement of Quality of Service in Cloud Computing

2021 ◽  
Author(s):  
Jianying Miao
Keyword(s):  
Cloud Computing ◽  
Quality Of Service ◽  
Task Scheduling ◽  
Virtual Machines ◽  
Cloud Environment ◽  
Simulation Environment ◽  
Attribute Matching ◽  
Resource Allocation Strategy ◽  
And Task

This thesis describes an innovative task scheduling and resource allocation strategy by using thresholds with attributes and amount (TAA) in order to improve the quality of service of cloud computing. In the strategy, attribute-oriented thresholds are set to decide on the acceptance of cloudlets (tasks), and the provisioning of accepted cloudlets on suitable resources represented by virtual machines (VMs,). Experiments are performed in a simulation environment created by Cloudsim that is modified for the experiments. Experimental results indicate that TAA can significantly improve attribute matching between cloudlets and VMs, with average execution time reduced by 30 to 50% compared to a typical non-filtering policy. Moreover, the tradeoff between acceptance rate and task delay, as well as between prioritized and non-prioritized cloudlets, may be adjusted as desired. The filtering type and range and the positioning of thresholds may also be adjusted so as to adapt to the dynamically changing cloud environment.

scholarly journals A Classification Method of the Incident By Extraction From Text

2019 ◽  
Vol 8 (12S2) ◽  
pp. 1-5 ◽  
Keyword(s):  
Early Warning ◽  
Classification Method ◽  
Accident Risk ◽  
Incident Case ◽  
Matching Rule ◽  
Attribute Matching ◽  
Feature Expression

In this research, in order to use for prediction of the accident risk which prevents serious accident and disaster, the method of detecting and classifying an incident from a text is proposed. A multi-attribute matching machine is used for detection and a classification. The feature expression is extracted from the incident case sentence currently released, and detection of an incident and the classification of an accident kind are carried out by the matching rule created from extraction data. Although classification precision was mostly as good as 0.783 as a result of the evaluation experiment, the room for an improvement for extraction precision was seen. The incident which was able to be managed with flawlessness or a slight injury although it was likely to get injured can warn of a big accident, and can urge evasion of it. Therefore, this research which leads to an early warning by detecting and classifying mechanically is meaningful. A future subject is an improvement of extraction precision.

A CORBA-Based Object Framework with Patient Identification Translation and Dynamic Linking

1999 ◽  
Vol 38 (01) ◽  
pp. 56-65 ◽  
Author(s):  
K. Ohe ◽  
C. Wang
Keyword(s):  
Patient Data ◽  
Heterogeneous Databases ◽  
Patient Identification ◽  
Web Based ◽  
Distributed Object Computing ◽  
Distributed Object ◽  
Attribute Matching ◽  
Computer Based ◽  
Dynamic Linking ◽  
Institutional Boundaries

Abstract:Exchanging and integration of patient data across heterogeneous databases and institutional boundaries offers many problems. We focused on two issues: (1) how to identify identical patients between different systems and institutions while lacking universal patient identifiers; and (2) how to link patient data across heterogeneous databases and institutional boundaries. To solve these problems, we created a patient identification (ID) translation model and a dynamic linking method in the Common Object Request Broker Architecture (CORBA) environment. The algorithm for the patient ID translation is based on patient attribute matching plus computer-based human checking; the method for dynamic linking is temporal mapping. By implementing these methods into computer systems with help of the distributed object computing technology, we built a prototype of a CORBA-based object framework in which the patient ID translation and dynamic linking methods were embedded. Our experiments with a Web-based user interface using the object framework and dynamic linking through the object framework were successful. These methods are important for exchanging and integrating patient data across heterogeneous databases and institutional boundaries.

Sign in / Sign up

Export Citation Format

Share Document