Efficient Algorithms for Dynamic Incomplete Decision Systems

Attribute reduction is a crucial problem in the process of data mining and knowledge discovery in big data. In incomplete decision systems, the model using tolerance rough set is fundamental to solve the problem by computing the redact to reduce the execution time. However, these proposals used the traditional filter approach so that the reduct was not optimal in the number of attributes and the accuracy of classification. The problem is critical in the dynamic incomplete decision systems which are more appropriate for real-world applications. Therefore, this paper proposes two novel incremental algorithms using the combination of filter and wrapper approach, namely IFWA_ADO and IFWA_DEO, respectively, for the dynamic incomplete decision systems. The IFWA_ADO computes reduct incrementally in cases of adding multiple objects while IFWA_DEO updates reduct when removing multiple objects. These algorithms are also verified on six data sets. Experimental results show that the filter-wrapper algorithms get higher performance than the other filter incremental algorithms.

Filter-Wrapper Incremental Algorithms for Finding Reduct in Incomplete Decision Systems When Adding and Deleting an Attribute Set

The tolerance rough set model is an effective tool to solve attribute reduction problem directly on incomplete decision systems without pre-processing missing values. In practical applications, incomplete decision systems are often changed and updated, especially in the case of adding or removing attributes. To solve the problem of finding reduct on dynamic incomplete decision systems, researchers have proposed many incremental algorithms to decrease execution time. However, the proposed incremental algorithms are mainly based on filter approach in which classification accuracy was calculated after the reduct has been obtained. As the results, these filter algorithms do not get the best result in term of the number of attributes in reduct and classification accuracy. This paper proposes two distance based filter-wrapper incremental algorithms: the algorithm IFWA_AA in case of adding attributes and the algorithm IFWA_DA in case of deleting attributes. Experimental results show that proposed filter-wrapper incremental algorithm IFWA_AA decreases significantly the number of attributes in reduct and improves classification accuracy compared to filter incremental algorithms such as UARA, IDRA.

Sentence Similarity Calculation Based on Probabilistic Tolerance Rough Sets

Sentence similarity calculation is one of the important foundations of natural language processing. The existing sentence similarity calculation measurements are based on either shallow semantics with the limitation of inadequately capturing latent semantics information or deep learning algorithms with the limitation of supervision. In this paper, we improve the traditional tolerance rough set model, with the advantages of lower time complexity and becoming incremental compared to the traditional one. And then we propose a sentence similarity computation model from the perspective of uncertainty of text data based on the probabilistic tolerance rough set model. It has the ability of mining latent semantics information and is unsupervised. Experiments on SICK2014 task and STSbenchmark dataset to calculate sentence similarity identify a significant and efficient performance of our model.

Pattern classification using tolerance rough sets based on non-additive grey relational analysis and DEMATEL

PurposeIt is challenging to derive an appropriate tolerance relation for tolerance rough set-based classifiers (TRSCs). The traditional tolerance rough set employs a simple distance function to determine the tolerance relation. However, such a simple function does not take into account criterion weights and the interaction among criteria. Further, the traditional tolerance relation ignores interdependencies concerning direct and indirect influences among patterns. This study aimed to incorporate interaction and interdependencies into the tolerance relation to develop non-additive grey TRSCs (NG-TRSCs).Design/methodology/approachFor pattern classification, this study applied non-additive grey relational analysis (GRA) and the decision-making trial and evaluation laboratory (DEMATEL) technique to solve problems arising from interaction and interdependencies, respectively.FindingsThe classification accuracy rates derived from the proposed NG-TRSC were compared to those of other TRSCs with distinctive features. The results showed that the proposed classifier was superior to the other TRSCs considered.Practical implicationsIn addition to pattern classification, the proposed non-additive grey DEMATEL can further benefit the applications for managerial decision-making because it simplifies the operations for decision-makers and enhances the applicability of DEMATEL.Originality/valueThis paper contributes to the field by proposing the non-additive grey tolerance rough set (NG-TRS) for pattern classification. The proposed NG-TRSC can be constructed by integrating the non-additive GRA with DEMATEL by using a genetic algorithm to determine the relevant parameters.

Designing Extreme Learning Machine Network Structure Based on Tolerance Rough Set

In this paper, we present a new extreme learning machine network structure on the basis of tolerance rough set. The purpose of this paper is to realize the high-efficiency and multi-dimensional ELM network structure. Various published algorithms have been applied to breast cancer datasets, but rough set is a fairly new intelligent technique that applies to predict breast cancer recurrence. We analyze Ljubljana Breast Cancer Dataset, firstly, obtain lower and upper approximations and calculate the accuracy and quality of the classification. The high values of the quality of classification and accuracy prove that the attributes selected can well approximate the classification. Rough sets approach is established to solve the prolem of tolerance.

About a Distance Measure and Application for Finding Reduct in Incomplete Decision Tables

Tolerance rough set model is an effective tool to reduce attributes in incomplete decision tables. Over 40 years, several attribute reduction methods have been proposed to improve the efficiency of execution time and the number of attributes of the reduct. However, they are classical filter algorithms, in which the classification accuracy of decision tables is computed after obtaining the reducts. Therefore, the obtained reducts of these algorithms are not optimal in terms of reduct cardinality and classification accuracy. In this paper, we propose a filter-wrapper algorithm to find a reduct in incomplete decision tables. We then use this measure to determine the importance of the property and select the attribute based on the calculated importance (filter phase). In the next step, we find the reduct with the highest classification accuracy by iterating over elements of the set containing the sequence of attributes selected in the first step (wrapper phase). To verify the effectiveness of the method, we conduct experiments on 6 famous UCI data sets. Experimental results show that the proposed method increase classification accuracy as well as reduce the cardinality of reduct compared to Algorithm 1 [12].

A DISTANCE BASED INCREMENTAL FILTER-WRAPPER ALGORITHM FOR FINDING REDUCT IN INCOMPLETE DECISION TABLES

Tolerance rough set model is an effective tool for attribute reduction in incomplete decision tables. In recent years, some incremental algorithms have been proposed to find reduct of dynamic incomplete decision tables in order to reduce computation time. However, they are classical filter algorithms, in which the classification accuracy of decision tables is computed after obtaining reduct. Therefore, the obtained reducts of these algorithms are not optimal on cardinality of reduct and classification accuracy. In this paper, we propose the incremental filter-wrapper algorithm IDS_IFW_AO to find one reduct of an incomplete desision table in case of adding multiple objects. The experimental results on some sample datasets show that the proposed filter-wrapper algorithm IDS_IFW_AO is more effective than the filter algorithm IARM-I [17] on classification accuracy and cardinality of reduct