An Ontology Based Frequent Itemset Method to Support Research Proposal Grouping for Research Project Selection

Abstract

Research proposal grouping is one of the most important tasks for research project selection in research funding agencies. In this paper, a novel ontology based frequent itemset method is proposed to deal with research proposal grouping problem. In the proposed method, a research ontology is firstly constructed to standardize research keywords. Secondly, frequent itemsets with different support degrees are extracted from research proposals based on research ontology. Thirdly, a new measure of similarity degree between two research proposals is developed and then a clustering algorithm is proposed to classify research proposals based on the similarity degree, in which some parameters are discussed, and the proper parameters are selected. Finally, when the number of research proposals in some clusters is still large, research proposals are further divided into small groups, in which the number of research proposals is approximately equal. The proposed method is validated based on the selection process at the National Natural Science Foundation of China (NSFC). The experimental results show that our proposed method can improve the efficiency and effectiveness of research proposal grouping, and is a potential and alternative one to support research project selection processes in other governments and private research funding agencies.

A comparative analysis of text representation, classification and clustering methods over real project proposals

“…In response to the need for automatic grouping of proposals, ontology-based text mining (OTMM) technology has frequently been used to automatically group project proposals (Ma et al ., 2012; Rathore et al ., 2013; Preethi and Lakshmi, 2013; Patil and Uddin, 2015; Xu et al ., 2013; Rajkamal, 2017). An ontology is a systematic description of a concept and its associated properties within a domain.…”

“…After developing a research ontology or performing feature extraction, documents are grouped according to their similarities. Various clustering techniques have been used for this purpose, such as the self-organizing map (SOM) (Wang et al ., 2015; Saravanan and Babu, 2017; Ma et al ., 2012; Ahmed et al ., 2020), expectation-maximization (Wang et al ., 2015; Rajkamal, 2017), k-means (Xu et al ., 2013; Patil and Uddin, 2015; Rajkamal, 2017), DBSCAN (Rajkamal, 2017) fuzzy c-means (Preethi and Lakshmi, 2013; Kumar et al ., 2019). Document classification methods include k-nearest neighbor (KNN) (Rathore et al ., 2013; Safi'ie et al ., 2018; Chowdhury and Schoen, 2020), naive bayes (Rathore et al ., 2013; Saravanan and Babu, 2017) and support vector machines (SVM) Saravanan and Babu (2017), Dominic and R (2021) and C4.5 (Rajkamal, 2017) are presented.…”

Analysis of practical applying of project selection method based on the possibility theory

“…Many of the proposed approaches are based on fuzzy logic. Examples can be found in [2][3][4][5][6][7][8][9][10][11].…”

“…In [4], an algorithm is proposed for selecting projects for a research agency, attention is focused on the objectives of research projects, while fuzzy mathematics is used to form an integral evaluation of projects. Work [5] is also devoted to a specific category of projects.…”