China
Africa
Explore chapters and articles related to this topic
Introduction
Published in Anuradha D. Thakare, Shilpa Laddha, Ambika Pawar, Hybrid Intelligent Systems for Information Retrieval, 2023
Anuradha D. Thakare, Shilpa Laddha, Ambika Pawar
Semantic search may quickly render appropriate information in less time due to ontology. Semantic similarity is calculated using WordNet. WordNet® is a massive lexical database in English. Nouns, verbs, adverbs, and adjectives are grouped into psychological analogs (synsets), each of which conveys a distinct notion. Synsets are linked by conceptual, semantic, and lexical relations. The application can be used to investigate the following structuring of definitively associated words and ideas. WordNet is also unreservedly and freely accessible for download. The structure of WordNet makes it a useful tool for natural language processing (NLP) and computational linguistics.
Semantic similarity research on case retrieval based on ontology
Published in Jimmy C.M. Kao, Wen-Pei Sung, Civil, Architecture and Environmental Engineering, 2017
Yiling Liu, Hong Duan, Lei Luo
The main contribution of this paper is proposal of a new semantic similarity measurement between cases based on ontology, called the Matched Genealogy Measurement (MGM), which is proved to perform well in matching human intuition. We took combat simulation as an example to illustrate ontology-based algorithm and described some typical combat simulation cases with the method as case library. The conclusion is supported by a user study and analysis of Window Distance. MGM could also meet different users’ requirements by adjusting coefficients. It is actually a general measurement of semantic similarity, and could be used widely in information retrieval and recommendation systems.
Similarity Principle—The Fundamental Principle of All Sciences
Published in Mark Chang, Artificial Intelligence for Drug Development, Precision Medicine, and Healthcare, 2020
Similarity (proximity, alikeness, affinity) can be measured in different ways. For example, one can use spatial distance, correlation, or comparison of local histograms, spectral properties, or even, with or without given patterns, a long data series, as in gene matching. Semantic similarity is a metric defined over a set of documents or terms, where the idea of distance between two of them is based on the likeness of their meaning or semantic content.
A hybrid approach based on linguistic analysis and fuzzy logic to ensure the surveillance of people having paranoid personality disorder towards Covid-19 on social media
Published in International Journal of General Systems, 2023
Mourad Ellouze, Seifeddine Mechti, Lamia Hadrich Belguith
The Universal Sentence Encoder (Cer et al. 2018) (USE) is a technique that consists of transforming the textual part into a set of numeric vectors. This technique can be used for many tasks like semantic similarity, clustering, text classification, and other natural language processing tasks. The training part of USE is available in Tensorflow-hub ,8 which is done by the combination of Transformer encoder (Hurtado 2020) and Deep Averaging Network (Gardner, Elhami, and Selmic 2019) (DAN). This combination allows it to have a higher accuracy and to be less expensive in terms of calculation time. In our work, we use this technique to convert the textual portion into a set of numeric vectors that facilitate the classification task afterwards. This technique allows us to solve many recognized problems related to the size of the corpus and the variety of terms since we are not obliged to do the training task. Moreover, this technique produces as results a set of standardized tables.
Big data analytics capabilities: a novel integrated fitness framework based on a tool-based content analysis
Published in Enterprise Information Systems, 2023
Sunil Pathak, Venkataraghavan Krishnaswamy, Mayank Sharma
We identified a total of 103 preliminary codes representing BDAC dimensions, but these codes/labels had semantic similarities in terms of nomenclature, indicating the need for synthesis. We used the axial coding approach to reduce the preliminary codes with similar meanings. This process was repeated by regrouping or renaming items into mutually exclusive dimensions. We used a Python program to perform semantic similarity checks using cosine similarity between words. Semantic similarity is based on word ontology and is used to assess the meaning of two pieces of text. In this work, we reconciled or renamed the codes with less than 100% (similar text) and greater than 25% cosine similarity (25% similarity in meaning). For the classification’s reliability assessment, two authors classified the codes independently, resulting in a 98.4% agreement on classification with a Cohen Kappa coefficient (k) value of .64, indicating substantial agreement (Richard and Koch 1977).
Disease prediction in data mining using association rule mining and keyword based clustering algorithms
Published in International Journal of Computers and Applications, 2020
To solve clustering, we need to define a similarity (or distance) between the objects. In agglomerative methods such as single link and complete link, similarity between individual objects is sufficient, but in partitioned clustering such as k-means and k-medoids cluster representative is also required to measure object-to-cluster similarity. Using semantic content, however, defining the representative of a cluster is not trivial. Fortunately, it is still possible to apply partitioned clustering even without the representatives. For example, an object can be assigned to such cluster that minimizes (or maximizes) the cost function where only the similarities between objects are needed. In this paper, we present a novel similarity measure between two sets of words, called matching similarity. We apply it to keyword-based clustering of services in location-based application. Assuming that we have measure for comparing semantic similarity of two words, the problem is to find a good measure to compare the sets of words. The proposed matching similarity solves the problem as follows. It iteratively pairs two most similar words between the objects and then repeats the process for the rest of the objects until one of the objects runs out of words. The remaining words are then matched just to their most similar counterpart in the other object.