Currently, there is a significant amount of public code in the IT communities, programming forums and code repositories. Many of these codes lack classification labels, or have imprecise labels, which causes inconvenience to code management and retrieval. Some classification methods have been proposed to automatically assign labels to the code. However, these methods mainly rely on code comments or surrounding text, and the classification effect is limited by the quality of them. So far, there are a few methods that rely solely on the code itself to assign labels to the code. In this paper, an encoder-only method is proposed to assign multiple labels to the code of an algorithmic problem, in which UniXcoder is employed to encode the input code and the encoding results correspond to the output labels through the classification heads. The proposed method relies only on the code itself. We construct a dataset to evaluate the proposed method, which consists of source code in three programming languages (C[Formula: see text], Java, Python) with a total size of approximately 120[Formula: see text]K. The results of the comparative experiment show that the proposed method has better performance in multi-label classification task of pure code than encoder–decoder methods.

Q-rung orthopair hesitant fuzzy set (q-ROHFS) is a potent and effective technique for dealing with more general and complex uncertainty. Multiple attribute decision-making (MADM) under complex uncertainty has been a key research issue. However in the existing MADM approaches, the fuzzy entropies involve much higher hesitancy degree loss and the fuzzy measure of attributes can not be determined objectively. Also these existing MADM methods under complex uncertainty have high data redundancy and low computational efficiency. In order to solve these problems, this paper proposes a novel q-rung orthopair hesitant fuzzy information MADM method based on the Choquet integral. Firstly, we give the axiomatic definition of q-rung orthopair hesitant fuzzy entropy (q-ROHFE) by extending dual hesitant fuzzy information entropy and derive the fuzzy entropy construction theorem and the two related q-ROHFE formulas, which greatly reduces the loss of hesitancy degree resulting from the existing fuzzy entropy. Secondly, combined with λfuzzy measure and proposed q-ROHFE, a constrained nonlinear fuzzy measure optimization model for q-rung orthopair hesitant fuzzy decision making is presented, which addresses the difficulty that existing research cannot determine the fuzzy measure of attributes under fuzzy MADM. Thirdly, an improved Choquet integral-based VIKOR approach based on the fuzzy measure computed by the model is developed. Finally, two real-life cases are shown to fully illustrate the suggested approach. Experiment results demonstrate that the proposed fuzzy entropy has much less hesitancy degree loss and the proposed approach significantly increases computational efficiency while reducing data redundancy. And our method has strong adaptability and scalability.

Parameter reduction is a crucial procedure for enhancing the effectiveness of decision-making processes in the theory of interval-valued fuzzy soft set, which is a developing and excellent mathematical tool for processing blurry data. However, there exist severe gaps in the existing literatures. That is, the existing parameter reduction algorithms for interval-valued fuzzy soft sets have high computational complexity and low success rate of finding reduction. By considering these gaps, this study firstly develops the notions of the mean degree of membership and Euclidean distance between parameters. And then a new Euclidean Distance-based parameter reduction algorithm for interval-valued fuzzy soft set is proposed. In addition, we compare our proposed algorithm with the existing parameter reduction algorithms in terms of computational complexity and success rate of finding reduction on 35 randomly generated data sets and a real-life application of Five-Star Hotels evaluation. Through the comparison results, it is clear that our approach has the lower computational complexity and higher success rate of finding reduction. Consequently, our method which is substantiated on the validity and superiority is a good solution for parameter reduction of interval-valued fuzzy soft sets.

In a typical formulation of decision-making under uncertainty, a decision-maker must choose a single-optimal option among many possible options. However, the problem of selecting a unique and optimal choice has remained a significant challenge to solve. In this article, we propose a new interval-valued intuitionistic fuzzy soft set (IVIFSS) based decision-making approach to address this problem. The proposed approach is based on the choice value and score value of membership/nonmembership degrees. Furthermore, three parameter reduction algorithms are proposed. We apply the proposed approaches on a real application to demonstrate their working and effectiveness. We also compare the proposed approach against the adjustable IVIFSSs approach and show that the proposed approach has lower computation overhead and enable a decision-maker to choose top options to make a proper decision.

There has been a rapid growth of interest in developing approaches that are capable of dealing with imprecision and uncertainty. To this end, an interval-valued fuzzy soft set (IVFSS) that combines soft set theory with interval-valued fuzzy set theory has been proposed to handle imprecision and uncertainty in applications such as decision-making problems. However, there has been little focus on parameter reduction of the interval-valued fuzzy soft sets, which is significant in decision-making problems. In this paper, we introduce four different definitions of parameter reduction in interval-valued fuzzy soft sets to satisfy different the needs of decision makers. We propose four heuristic algorithms of parameter reduction. Finally, the algorithms are compared and summarized from the aspects of easy degree of finding reduction, applicability, reduction result, exact level for reduction, multiusability, applied situation, and computational complexity. The results of the experiment show that the methods reduce the redundant parameters while preserving certain decision abilities.