Item semantic annotation method based on use event of heterogeneous item

Abstract

The invention relates to an item semantic annotation method based on a use event of a heterogeneous item. The method comprises the following steps: step 1) of establishing an item potential relationship intensity mining model; step 2) of extracting an item implicit feature; step 3) of extracting an item explicit feature; step 4) of training an item semantic annotation model; after the above four features are extracted, combining (FLatent+Ftext+FS+FT) together as training data of a binary classification SVM classifier, and gathering labels predicted by all independent binary classification device as the semantic annotation result of the item. The item semantic annotation method has the advantages that based on a latent variable model establishing item implicit relationship strength based onthe use event of the heterogeneous item, the item implicit feature is extracted therefrom, the category semantic label of the given item is predicted by comprehensively utilizing the explicit and implicit features, and the method is significantly superior to an existing item labeling method based on explicit features in terms of efficiency and performance and the like.

Description

technical field [0001] The present invention relates to an item semantic tagging method, which mainly establishes a latent variable model of item implicit relationship strength based on heterogeneous item usage events, extracts item implicit features from it, and comprehensively utilizes the display and implicit features to predict the category semantics of a given item Label. Background technique [0002] In the cyber-physical intelligent space, people can search for items, identify and browse items, and carry out semantic annotation on items, which is a key step to realize the above-mentioned cyber-physical space applications. However, the traditional semantic annotation technology faces severe challenges in the cyber-physical space: due to the fact that the text data of items in the cyber-physical space is usually few and confusing noises, text-based annotation technology {[1] WANG D, ZHANG H, LIU R ,et al.Feature selection based on term frequency and T-test for text cat...

AI Technical Summary

Problems solved by technology

However, the traditional semantic annotation technology faces severe challenges in the cyber-physical space: due to the fact that the text data of items in the cyber-physical space is usually few and confusing noises, text-based annotation technology {[1] WANG D, ZHANG H, LIU R ,et al.Feature selection based on term frequency and T-test for text categorization[C] / / Proceedings of 21stACM International Conference on Information and Knowledge Management,CIKM'12,Maui,HI,USA,October 29-November 02,2012: 1482–1486.http: / / doi.acm.org / 10.1145 / 2396761.2398457.[2]XUE X,ZHOU Z.Distributional Features for TextCategorization[J].IEEE Trans.Knowl.Data Eng.,2009,21(3) :428–442.https: / / doi.org / 10.1109 / TKDE.2008.166.[3]XU Z,JIN R,HUANG K,et al.Semi-supervisedtext categorization by active search[C] / / Proceedings of the 17th ACMConference on Information and Knowledge Management, CIKM 2008, Napa Valley, California, USA, October 26-30, 2008:1517–1518. http: / / doi.acm.org / 10.1145 / 1458082.1458364.} Unable to achieve satisfactory labeling Performance; there is usually a lack of a large number of labeled items in the network-physical space as a training set, and manual labeling is time-consuming and laborious; due to the need for time-consuming and laborious prior knowledge and the construction of item semantic network models, such as manually defined in a unified format (such as a resource description framework) The description of the item and its corresponding concept make the item annotation method based on the semantic web {[4] COCCOLI M, TORRE I.Interaction with Objects andObjects Annotation in the Semantic Web of Things[C] / / Proceedings of The 20th International Conference on Distributed Multimedia Systems: Research papers on distributed multimedia systems, distance education technologies and visual languages ​​and computing, Pittsburgh, PA, USA, August 27-29, 2014:383–390.[5] DE S, BARNAGHI P, BAUER M, et al. Service modeling for the Internet of Things[C] / / Proceedings 2011 Federated Conference on Computer Science and Information Systems (FedCSIS), 2011:949–955.[6] ATANASOV I, NIKOLOV A ,PENCHEVA E,et al.An Approach to Data Annotation for Internet of Things[J].IJITWE,2015,10(4):1–19. https: / / doi.org / 10.4018 / IJITWE .2015100101.[7]NEVAT I, PETERS G W, AVNIT K, et al.Location of Things: Geospatial Tagging for IoT Using Time-of-Arrival[J].IEEE Trans.Signal and Information Processing over Networks,2016,2(2 ):174–185.https: / / doi.org / 10.1109 / TSIPN.2016.2531422.} does not have scalability; because the items in the network-physical space do not have display links similar to web page URLs or friend relationships in social networks, so Link-based annotation method {[8]MCDOWELL L K. Relational active learning for link-based classification[C] / / Proceedings of 2015 IEEE International Conference on DataScience and Advanced Analytics, DSAA 2015, Campus des Cordeliers, Paris, France, October 19 -21,2015:1–10.https: / / doi.org / 10.1109 / DSAA.2015.7344798.[9]TIAN Y,YANG Q,HUANG T,et al.Learning Contextual Dependency Network Models for Link-Based Classification[J ].IEEE Trans.Knowl.Data Eng.,2006,18(11):1482–1496.https: / / doi.org / 10.1109 / TKDE.2006.178.[10]LIMAYE G,SARAWAGI S,CHAKRABARTI S.Annotating and Searching Web Tables Using Entities,Types and Relationships[J].PVLDB,2010,3(1):1338–1347.http: / / www.comp.nus.edu.sg / ~vldb2010 / proceedings / files / papers / R118.pdf .} does not apply to item annotations in cyber-physical space

Images