Schema-summarization in linked-data-based feature selection for recommender systems (original) (raw)

Schema-aware Feature Selection in Linked Data-based Recommender Systems

2017

Semantics-aware recommendation engines have emerged as a new family of systems able to exploit the semantics encoded in unstructured and structured information sources to provide better results in terms of accuracy, diversity and novelty as well as to foster the provisioning of new services such as explanation. In the rising of these new recommender systems, an important role has been played by Linked Data (LD). However, as Linked Data is often very rich and contains many information that may result irrelevant and noisy, an initial step of feature selection may be required in order to select the most meaningful portion of the original dataset. Many approaches have been proposed in the literature for feature selection that exploit different statistical dimensions of the original data. In this paper we investigate the role of the semantics encoded in an ontological hierarchy via schema-summarization when exploited to select the most relevant properties for a recommendation task.

Automatic Selection of Linked Open Data Features in Graph-based Recommender Systems

2015

In this paper we compare several techniques to automatically feed a graph-based recommender system with features extracted from the Linked Open Data (LOD) cloud. Specifically, we investigated whether the integration of LOD-based features can improve the e↵ectiveness of a graph-based recommender system and to what extent the choice of the features selection technique can influence the behavior of the algorithm by endogenously inducing a higher accuracy or a higher diversity. The experimental evaluation showed a clear correlation between the choice of the feature selection technique and the ability of the algorithm to maximize a specific evaluation metric. Moreover, our algorithm fed with LODbased features was able to overcome several state-of-the-art baselines: this confirmed the e↵ectiveness of our approach and suggested to further investigate this research line.

A Linked Dataverse Knows Better: Boosting Recommendation Quality Using Semantic Knowledge

2011

Abstract: The advent of Linked Open Data (LOD) gave birth to a plethora of open datasets freely available to everyone. Accompanied with LOD, a new research field arises focusing on how to handle and to take advantage of this huge amount of data. In this paper, we introduce a novel approach utilizing and aggregating open datasets to compute the most-related entities for a set of weighted input entities.

Weighted Linked Data based Unsupervised Feature Selection

International Conference on Science and Innovative Engineering, 2015

a large volume of dataset consists of more number of features. Among all the features which feature provides more accuracy in prediction is unknown. Decision support system performance is based on machine learning techniques such as clustering and classification. Objective of the system is to select minimum features and provide more accuracy in prediction. Linked data provides more information comparatively attribute value data. To achieve this task, process the attribute value data and prepare linked data using graph regularization and social dimension reduction. By using Link weight prediction techniques exploit the strong correlation between the links and reduce the noisy data. Attractiveness-based community detection (ABCD) algorithm used to extract and select the best community in the weighted graph for Link prediction. Based on the best clusters, the optimal feature set is selected. Accuracy of the clusters in this system is increased, while compared to the performance of existing system.

Executing, Comparing, and Reusing Linked-Data-Based Recommendation Algorithms With the Allied Framework

Semantic Web Science and Real-World Applications

Data published on the web following the principles of linked data has resulted in a global data space called the Web of Data. These principles led to semantically interlink and connect different resources at data level regardless their structure, authoring, location, etc. The tremendous and continuous growth of the Web of Data also implies that now it is more likely to find resources that describe real-life concepts. However, discovering and recommending relevant related resources is still an open research area. This chapter studies recommender systems that use linked data as a source containing a significant amount of available resources and their relationships useful to produce recommendations. Furthermore, it also presents a framework to deploy and execute state-of-the-art algorithms for linked data that have been re-implemented to measure and benchmark them in different application domains and without being bound to a unique dataset.

LODify : A Hybrid Recommender System based on Linked Open Data

2014

We propose LODify, a hybrid recommendation method which measures the semantic similarity of items or resources of interest and combines this with user ratings to make recommendations across diverse domains. The semantic similarity metric draws on information theory and computes the similarity of items based on the information content of their shared characteristics. Detailed semantic analysis of items, considering the special characteristics of Linked Data represented using various kinds of relations, incorporating the relative importance of each relation into the similarity measurement, and successful handling of the item cold-start problem are among the key benefits of the presented approach. We demonstrate how this approach can be successfully applied to provide recommendations and to predict user ratings. 1 LODify and its Innovation Linking Open Data (LOD) project is one of the successful initiatives of the Web of Data which aims to publish and link public datasets in a wide var...