Pub Date : 2023-10-01DOI: 10.1142/s021848852350037x
Qiuyan Zhan, Lesheng Jin, Ronald R. Yager
In real life, multiple attribute decision problems (MADM) can be applied in different areas and numerous related extensions and methodologies have been proposed by researchers. Combining three-way TOPSIS decision ideas with MADM is a feasible and meaningful research direction. In light of this, this paper generalizes the classical TOPSIS method with the help of mean and standard deviation and proposes the so-called modified three-way TOPSIS. First, using a pair of thresholds which is derived by mean and standard deviation, we divide decision alternatives into three segments, and then a preliminary rank results of decision alternatives can be obtained. Furthermore, in each decision region, we use two ranking regulations (one-way TOPSIS or modified two-way TOPSIS method) to rank decision alternatives. A practical example of urban expressway route selection illustrates the feasibility of the proposed method. Finally, we test the feasibility and validity of the modified three-way TOPSIS method by comparing with some existing method.
{"title":"A Modified TOPSIS Approach with Three-Way Decision","authors":"Qiuyan Zhan, Lesheng Jin, Ronald R. Yager","doi":"10.1142/s021848852350037x","DOIUrl":"https://doi.org/10.1142/s021848852350037x","url":null,"abstract":"In real life, multiple attribute decision problems (MADM) can be applied in different areas and numerous related extensions and methodologies have been proposed by researchers. Combining three-way TOPSIS decision ideas with MADM is a feasible and meaningful research direction. In light of this, this paper generalizes the classical TOPSIS method with the help of mean and standard deviation and proposes the so-called modified three-way TOPSIS. First, using a pair of thresholds which is derived by mean and standard deviation, we divide decision alternatives into three segments, and then a preliminary rank results of decision alternatives can be obtained. Furthermore, in each decision region, we use two ranking regulations (one-way TOPSIS or modified two-way TOPSIS method) to rank decision alternatives. A practical example of urban expressway route selection illustrates the feasibility of the proposed method. Finally, we test the feasibility and validity of the modified three-way TOPSIS method by comparing with some existing method.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136152384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-01DOI: 10.1142/s021848852397005x
Vladik Kreinovich
International Journal of Uncertainty, Fuzziness and Knowledge-Based SystemsVol. 31, No. 05, pp. 889-890 (2023) No AccessInterval Methods in Knowledge RepresentationVladik KreinovichVladik KreinovichDepartment of Computer Science, University of Texas at El Paso, El Paso, TX 79968, USAhttps://doi.org/10.1142/S021848852397005XCited by:0 (Source: Crossref) Previous AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsRecommend to Library ShareShare onFacebookTwitterLinked InRedditEmail Remember to check out the Most Cited Articles! Check out our titles on Fuzzy Logic & Z-Numbers With a wide range of areas, you're bound to find something you like. FiguresReferencesRelatedDetails Recommended Vol. 31, No. 05 Metrics History PDF download
国际不确定性，模糊性和基于知识的系统杂志vol . 3。31, No. 05, pp. 889-890 (2023) No AccessInterval Methods in Knowledge表示弗拉迪克·克雷诺维奇弗拉迪克·克雷诺维奇计算机科学系，德克萨斯大学埃尔帕索，得克萨斯州埃尔帕索79968，美国埃尔帕索https://doi.org/10.1142/S021848852397005XCited by:0(来源:交叉参考)Previous AboutSectionsPDF/EPUB tools添加到收藏夹下载CitationsTrack citations推荐到图书馆分享分享在facebook上推特链接在redditemail记得查看被引用最多的文章!看看我们的标题模糊逻辑和z -数字与广泛的领域，你一定会找到你喜欢的东西。FiguresReferencesRelatedDetails推荐卷31,No. 05指标历史PDF下载
{"title":"Interval Methods in Knowledge Representation","authors":"Vladik Kreinovich","doi":"10.1142/s021848852397005x","DOIUrl":"https://doi.org/10.1142/s021848852397005x","url":null,"abstract":"International Journal of Uncertainty, Fuzziness and Knowledge-Based SystemsVol. 31, No. 05, pp. 889-890 (2023) No AccessInterval Methods in Knowledge RepresentationVladik KreinovichVladik KreinovichDepartment of Computer Science, University of Texas at El Paso, El Paso, TX 79968, USAhttps://doi.org/10.1142/S021848852397005XCited by:0 (Source: Crossref) Previous AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsRecommend to Library ShareShare onFacebookTwitterLinked InRedditEmail Remember to check out the Most Cited Articles! Check out our titles on Fuzzy Logic & Z-Numbers With a wide range of areas, you're bound to find something you like. FiguresReferencesRelatedDetails Recommended Vol. 31, No. 05 Metrics History PDF download","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136129890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-01DOI: 10.1142/s0218488523500356
Gugulothu Venkanna, K. F. Bharati
This paper intends to develop a novel framework for text document clustering with the aid of a new improved meta-heuristic algorithm. Initially, the features are selected from the text document by subjecting each word under Term Frequency-Inverse Document Frequency (TF-IDF) computation. Subsequently, centroid selection plays a vital role in cluster formation, which is done using a new Improved Lion Algorithm (LA) termed as Cross over probability-based LA model (CP-LA). As a novelty, this paper introduced a new inter and intracluster similarity model. Moreover, this centroid selection is made in such a way that the proposed adaptive weighted similarity should be minimal. Based on the characteristics of the document, the weights are automatically adapted with the similarity measure. The proposed adaptive weighted similarity function involves the inter-cluster, and intra-cluster similarity of both ordered and unordered documents. Finally, the superiority of the proposed over other models is proved under different performance measures.
{"title":"Improved Meta-Heuristic Model for Text Document Clustering by Adaptive Weighted Similarity","authors":"Gugulothu Venkanna, K. F. Bharati","doi":"10.1142/s0218488523500356","DOIUrl":"https://doi.org/10.1142/s0218488523500356","url":null,"abstract":"This paper intends to develop a novel framework for text document clustering with the aid of a new improved meta-heuristic algorithm. Initially, the features are selected from the text document by subjecting each word under Term Frequency-Inverse Document Frequency (TF-IDF) computation. Subsequently, centroid selection plays a vital role in cluster formation, which is done using a new Improved Lion Algorithm (LA) termed as Cross over probability-based LA model (CP-LA). As a novelty, this paper introduced a new inter and intracluster similarity model. Moreover, this centroid selection is made in such a way that the proposed adaptive weighted similarity should be minimal. Based on the characteristics of the document, the weights are automatically adapted with the similarity measure. The proposed adaptive weighted similarity function involves the inter-cluster, and intra-cluster similarity of both ordered and unordered documents. Finally, the superiority of the proposed over other models is proved under different performance measures.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136129107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-01DOI: 10.1142/s0218488523500381
Le Thanh Tung, Dang Hoang Tam
This paper is intended to investigate fuzzy variational problems of several dependent variables. Firstly, we establish both necessary and sufficient optimality conditions for fundamental fuzzy variational problems and fuzzy variational problems with natural boundary conditions. Then, the necessary and sufficient optimality conditions for fuzzy variational problems with isoperimetric constraints and holonomic constraints are discussed. Some examples are given to illustrate our results.
{"title":"Necessary and Sufficient Optimality Conditions for Fuzzy Variational Problems of Several Dependent Variables in Terms of Fuzzy Granular Derivatives","authors":"Le Thanh Tung, Dang Hoang Tam","doi":"10.1142/s0218488523500381","DOIUrl":"https://doi.org/10.1142/s0218488523500381","url":null,"abstract":"This paper is intended to investigate fuzzy variational problems of several dependent variables. Firstly, we establish both necessary and sufficient optimality conditions for fundamental fuzzy variational problems and fuzzy variational problems with natural boundary conditions. Then, the necessary and sufficient optimality conditions for fuzzy variational problems with isoperimetric constraints and holonomic constraints are discussed. Some examples are given to illustrate our results.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136152385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500289
Jie Ren, Ping Zhu
Uncertainty measures are instrumental in describing the classification abilities in information systems, and uncertain information has been measured and processed with granular computing theory. While the fuzzy set-valued information system is a generalization of fuzzy information systems, the relationship between the information granulation and the uncertainty in fuzzy set-valued information systems remains to be studied. This paper probes into uncertainty measures in fuzzy set-valued information systems based on the fuzzy [Formula: see text]-neighborhood and the idea of granulation. Specifically, the fuzzy [Formula: see text]-neighborhood similarity relation that reflects the similarity between two objects is defined in terms of the nearness degree. We propose the concepts of information granules and granular structures induced by fuzzy [Formula: see text]-neighborhood similarity relations, based on which we introduce the granularity measures and rough approximation measures of granular structures in fuzzy set-valued information systems. Given the situation of decision information systems, we propose the granularity-based rough approximation measures by combining granularity measures with rough approximation measures. Experiment results and effectiveness analysis show that the measures we proposed are reasonable and feasible.
{"title":"Uncertainty Measures in Fuzzy Set-Valued Information Systems Based on Fuzzy β-Neighborhood Similarity Relations","authors":"Jie Ren, Ping Zhu","doi":"10.1142/s0218488523500289","DOIUrl":"https://doi.org/10.1142/s0218488523500289","url":null,"abstract":"Uncertainty measures are instrumental in describing the classification abilities in information systems, and uncertain information has been measured and processed with granular computing theory. While the fuzzy set-valued information system is a generalization of fuzzy information systems, the relationship between the information granulation and the uncertainty in fuzzy set-valued information systems remains to be studied. This paper probes into uncertainty measures in fuzzy set-valued information systems based on the fuzzy [Formula: see text]-neighborhood and the idea of granulation. Specifically, the fuzzy [Formula: see text]-neighborhood similarity relation that reflects the similarity between two objects is defined in terms of the nearness degree. We propose the concepts of information granules and granular structures induced by fuzzy [Formula: see text]-neighborhood similarity relations, based on which we introduce the granularity measures and rough approximation measures of granular structures in fuzzy set-valued information systems. Given the situation of decision information systems, we propose the granularity-based rough approximation measures by combining granularity measures with rough approximation measures. Experiment results and effectiveness analysis show that the measures we proposed are reasonable and feasible.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"48 36","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72390591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500277
R. Savaş
The purpose of this article is to present the concepts of double statistical convergence in credibility and [Formula: see text]-double statistical convergence in credibility in pringsheim sense. By using these definitions we present a natural multidimensional extension of Credibility theory via Summability methods.
{"title":"Multidimensional Statistical Convergence in Credibility Theory","authors":"R. Savaş","doi":"10.1142/s0218488523500277","DOIUrl":"https://doi.org/10.1142/s0218488523500277","url":null,"abstract":"The purpose of this article is to present the concepts of double statistical convergence in credibility and [Formula: see text]-double statistical convergence in credibility in pringsheim sense. By using these definitions we present a natural multidimensional extension of Credibility theory via Summability methods.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"103 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79944784","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500253
Huayan Wen, Xinxing Wu, G. Çayli
Uninorms, as important generalizations of triangular norms and conorms, let the identity [Formula: see text] exist anywhere on a bounded lattice. In this paper, we focus on new characterizations of uninorms allowed to act on more general bounded lattices. In particular, we present several necessary and sufficient conditions to verify the construction approaches introduced by (Çaylı and Karaçal, Kybernetika 53 (2017) 394–417) and (Çaylı, Fuzzy Sets Syst. 395 (2020) 107–129), yielding a uninorm on bounded lattices.
{"title":"Characterizing Some Types of Uninorms on Bounded Lattices","authors":"Huayan Wen, Xinxing Wu, G. Çayli","doi":"10.1142/s0218488523500253","DOIUrl":"https://doi.org/10.1142/s0218488523500253","url":null,"abstract":"Uninorms, as important generalizations of triangular norms and conorms, let the identity [Formula: see text] exist anywhere on a bounded lattice. In this paper, we focus on new characterizations of uninorms allowed to act on more general bounded lattices. In particular, we present several necessary and sufficient conditions to verify the construction approaches introduced by (Çaylı and Karaçal, Kybernetika 53 (2017) 394–417) and (Çaylı, Fuzzy Sets Syst. 395 (2020) 107–129), yielding a uninorm on bounded lattices.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"7 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83216799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500307
Z. Y. Chen, Ruei-yuan Wang, Y. Meng, Timothy Chen
Based on this article, a fuzzy NN (neural network) based on the EBA (evolved bat algorithm) was developed to devise adaptive control with gray signal prediction to provide asymptomatic stability and increased driving comfort. The method is used to assess plant nonlinearity and to perform structural tracking of the signal. The set of Gray’s differential equations is applied to Gray’s model (GM) (n, h), which has been an active system model. In the model, n is the order of the Gray’s differential equation and h is the number of variables considered. In this paper, a GM(2.1) has been utilised to achieve advanced nonlinear motion of a system, allowing the controller to demonstrate the efficiency and stability of the whole system in a Lyapunov-like expression. The controller design standard for a MEW (mechanical elastic wheel) is presented, creating a realistic framework in mathematical for practical engineering applications.
{"title":"Modeling Control and Forecasting Nonlinear Systems Based on Grey Signal Theory","authors":"Z. Y. Chen, Ruei-yuan Wang, Y. Meng, Timothy Chen","doi":"10.1142/s0218488523500307","DOIUrl":"https://doi.org/10.1142/s0218488523500307","url":null,"abstract":"Based on this article, a fuzzy NN (neural network) based on the EBA (evolved bat algorithm) was developed to devise adaptive control with gray signal prediction to provide asymptomatic stability and increased driving comfort. The method is used to assess plant nonlinearity and to perform structural tracking of the signal. The set of Gray’s differential equations is applied to Gray’s model (GM) (n, h), which has been an active system model. In the model, n is the order of the Gray’s differential equation and h is the number of variables considered. In this paper, a GM(2.1) has been utilised to achieve advanced nonlinear motion of a system, allowing the controller to demonstrate the efficiency and stability of the whole system in a Lyapunov-like expression. The controller design standard for a MEW (mechanical elastic wheel) is presented, creating a realistic framework in mathematical for practical engineering applications.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"20 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91266193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500290
Ke Lei, Yonghong Wu
As a novel intelligent algorithm, fireworks algorithm (FWA) is applied to deal with different types of optimization problems. Since FWA’s search processes are relatively simple, it is inefficient. In this paper, a dynamic multi-swarm competitive fireworks algorithm (DMCFWA) is developed to enhance the search capability of FWA. Firstly, based on the scaling coefficient updated by utilizing the fitness value of the optimal firework, the dynamic explosion amplitude strategy is proposed to improve the search capability of the best firework. Secondly, utilizing the location information of the fireworks, an improved search method is designed to enhance the local search capability of firework swarms. Thirdly, a multi-swarm independent selection technique and a restart operation are adopted to boost its abilities of global exploration and local exploitation. Finally, to reduce the computational cost of FWA, a new initialization method is used and a new model for calculating the spark number is embedded in DMCFWA. By adopting these strategies, DMCFWA easily implements and does well in exploitation and exploration. CEC2017 test suite and four engineering constraint problems are used to demonstrate the performance of DMCFWA. Experimental results show that DMCFWA performs more effectively and stably than its competitors.
{"title":"Dynamic Multi-Swarm Competitive Fireworks Algorithm for Global Optimization and Engineering Constraint Problems","authors":"Ke Lei, Yonghong Wu","doi":"10.1142/s0218488523500290","DOIUrl":"https://doi.org/10.1142/s0218488523500290","url":null,"abstract":"As a novel intelligent algorithm, fireworks algorithm (FWA) is applied to deal with different types of optimization problems. Since FWA’s search processes are relatively simple, it is inefficient. In this paper, a dynamic multi-swarm competitive fireworks algorithm (DMCFWA) is developed to enhance the search capability of FWA. Firstly, based on the scaling coefficient updated by utilizing the fitness value of the optimal firework, the dynamic explosion amplitude strategy is proposed to improve the search capability of the best firework. Secondly, utilizing the location information of the fireworks, an improved search method is designed to enhance the local search capability of firework swarms. Thirdly, a multi-swarm independent selection technique and a restart operation are adopted to boost its abilities of global exploration and local exploitation. Finally, to reduce the computational cost of FWA, a new initialization method is used and a new model for calculating the spark number is embedded in DMCFWA. By adopting these strategies, DMCFWA easily implements and does well in exploitation and exploration. CEC2017 test suite and four engineering constraint problems are used to demonstrate the performance of DMCFWA. Experimental results show that DMCFWA performs more effectively and stably than its competitors.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"64 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75704376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1142/s0218488523500319
Yue Gao, Jian-Wei Liu
Nowadays, many sequences of events are generated in areas as diverse as healthcare, finance, and social network. People have been studying these data for a long time. They hope to predict the type and occurrence time of the next event by using relationships among events in the data. recently, with the successful application of Recurrent Neural Network (RNN) in natural language processing, it has been introduced into point process. However, RNN cannot capture the long-term dependence among events well, and self-attention can partially mitigate this problem precisely. Transformer Hawkes Process (THP) using self-attention greatly improves the performance of the Hawkes Process, but THP cannot ignore the effect of irrelevant events, which will affect the computational complexity and prediction accuracy of the model. In this paper, we propose an Adaptively Sparse Transformers Hawkes Process (ASTHP). ASTHP considers the periodicity and nonlinearity of event time in the time encoding process. The sparsity of the ASTHP is achieved by substituting Softmax with [Formula: see text]-entmax: [Formula: see text]-entmax is a differentiable generalization of Softmax that allows unrelated events to gain exact zero weight. By optimizing the neural network parameters, different attention heads can adaptively select sparse modes (from Softmax to Sparsemax). Compared with the existing models, ASTHP model not only ensures the prediction performance but also improves the interpretability of the model. For example, the accuracy of ASTHP model on MIMIC-II dataset is improved by nearly 3 percentage points, and the model fitting degree and stability are also improved significantly.
{"title":"Adaptively Sparse Transformers Hawkes Process","authors":"Yue Gao, Jian-Wei Liu","doi":"10.1142/s0218488523500319","DOIUrl":"https://doi.org/10.1142/s0218488523500319","url":null,"abstract":"Nowadays, many sequences of events are generated in areas as diverse as healthcare, finance, and social network. People have been studying these data for a long time. They hope to predict the type and occurrence time of the next event by using relationships among events in the data. recently, with the successful application of Recurrent Neural Network (RNN) in natural language processing, it has been introduced into point process. However, RNN cannot capture the long-term dependence among events well, and self-attention can partially mitigate this problem precisely. Transformer Hawkes Process (THP) using self-attention greatly improves the performance of the Hawkes Process, but THP cannot ignore the effect of irrelevant events, which will affect the computational complexity and prediction accuracy of the model. In this paper, we propose an Adaptively Sparse Transformers Hawkes Process (ASTHP). ASTHP considers the periodicity and nonlinearity of event time in the time encoding process. The sparsity of the ASTHP is achieved by substituting Softmax with [Formula: see text]-entmax: [Formula: see text]-entmax is a differentiable generalization of Softmax that allows unrelated events to gain exact zero weight. By optimizing the neural network parameters, different attention heads can adaptively select sparse modes (from Softmax to Sparsemax). Compared with the existing models, ASTHP model not only ensures the prediction performance but also improves the interpretability of the model. For example, the accuracy of ASTHP model on MIMIC-II dataset is improved by nearly 3 percentage points, and the model fitting degree and stability are also improved significantly.","PeriodicalId":50283,"journal":{"name":"International Journal of Uncertainty Fuzziness and Knowledge-Based Systems","volume":"19 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75156014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：