Purpose

This work provides a generic review of the existing data mining ontologies (DMOs) and also provides a base platform for ontology developers and researchers for gauging the ontologies for satisfactory coverage and usage.

Design/methodology/approach

The study uses a systematic literature review approach to identify 35 DMOs in the domain between the years 2003 and 2021. Various parameters, like purpose, design methodology, operations used, language representation, etc. are available in the literature to review ontologies. Accompanying the existing parameters, a few parameters, like semantic reasoner used, knowledge representation formalism was added and a list of 20 parameters was prepared. It was then segregated into two groups as generic parameters and core parameters to review DMOs.

Findings

It was observed that among the 35 papers under the study, 26 papers were published between the years 2006 and 2016. Larisa Soldatova, Saso Dzeroski and Pance Panov were the most productive authors of these DMO-related publications. The ontological review indicated that most of the DMOs were domain and task ontologies. Majority of ontologies were formal, modular and represented using web ontology language (OWL). The data revealed that Ontology development 101, METHONTOLOGY was the preferred design methodology, and application-based approaches were preferred for evaluation. It was also observed that around eight ontologies were accessible, and among them, three were available in ontology libraries as well. The most reused ontologies were OntoDM, BFO, OBO-RO, OBI, IAO, OntoDT, SWO and DMOP. The most preferred ontology editor was Protégé, whereas the most used semantic reasoner was Pellet. Even ontology metrics for 16 DMOs were also available.

Originality/value

This paper carries out a basic level review of DMOs employing a parametric approach, which makes this study the first of a kind for the review of DMOs.

中文翻译：

---

数据挖掘本体综述

目的

这项工作提供了对现有数据挖掘本体 (DMO) 的一般审查，还为本体开发人员和研究人员提供了一个基础平台，用于衡量本体的满意覆盖率和使用情况。

设计/方法/方法

该研究使用系统的文献回顾方法来确定 2003 年至 2021 年期间该领域的 35 个 DMO。文献中提供了各种参数，如目的、设计方法、使用的操作、语言表示等来回顾本体。在现有参数的基础上，添加了一些参数，如使用的语义推理器、知识表示形式主义，并准备了 20 个参数的列表。然后将其分为两组作为通用参数和核心参数来审查 DMO。

发现

据观察，在该研究的 35 篇论文中，有 26 篇是在 2006 年至 2016 年期间发表的。Larisa Soldatova、Saso Dzeroski 和 Pance Panov 是这些 DMO 相关出版物的最高产作者。本体审查表明，大多数 DMO 是领域和任务本体。大多数本体是正式的、模块化的，并使用 Web 本体语言 (OWL) 表示。数据显示，本体开发 101，方法论是首选的设计方法，基于应用的方法是首选的评估方法。还观察到大约有八个本体可以访问，其中三个在本体库中也是可用的。重用最多的本体是 OntoDM、BFO、OBO-RO、OBI、IAO、OntoDT、SWO 和 DMOP。最受欢迎的本体编辑器是 Protégé，而最常用的语义推理器是 Pellet。甚至 16 个 DMO 的本体度量也可用。

原创性/价值

本文采用参数化方法对 DMOs 进行了基本层面的审查，这使得本研究成为第一个对 DMOs 进行审查的研究。