Pub Date : 2021-07-12DOI: 10.1080/0952813X.2021.1952654
Walid Bannour, A. Maalel, H. Ghézala
ABSTRACT With the frequent occurrence of natural and man-made disasters, emergency management has become an active research field aiming at saving lives and reducing environmental and economic losses. Due to the complexity of crisis situations, emergency managers need to be assisted in making critical and effective decisions. Case-based reasoning (CBR) methodology has been widely adopted to support emergency decision makers in their tasks. This paper presents a comprehensive literature review of recent emergency management CBR systems reported in peer-reviewed journals and ICCBR conference proceedings between 2000 and 2020. Recent development trends of emergency management CBR systems are identified in terms of their purposes, application contexts and techniques used for their development. Finally, opportunities to improve emergency management CBR systems are outlined.
{"title":"Emergency Management Case-Based Reasoning Systems: A Survey of Recent Developments","authors":"Walid Bannour, A. Maalel, H. Ghézala","doi":"10.1080/0952813X.2021.1952654","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1952654","url":null,"abstract":"ABSTRACT With the frequent occurrence of natural and man-made disasters, emergency management has become an active research field aiming at saving lives and reducing environmental and economic losses. Due to the complexity of crisis situations, emergency managers need to be assisted in making critical and effective decisions. Case-based reasoning (CBR) methodology has been widely adopted to support emergency decision makers in their tasks. This paper presents a comprehensive literature review of recent emergency management CBR systems reported in peer-reviewed journals and ICCBR conference proceedings between 2000 and 2020. Recent development trends of emergency management CBR systems are identified in terms of their purposes, application contexts and techniques used for their development. Finally, opportunities to improve emergency management CBR systems are outlined.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"23 1","pages":"35 - 58"},"PeriodicalIF":2.2,"publicationDate":"2021-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85079754","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 : 2021-07-10DOI: 10.1080/0952813X.2021.1949753
Guanghao Jin, Qingzeng Song
ABSTRACT Currently, deep learning methods have been widely applied to many fields like classification. Generally, these methods use the technology like transferring to make a model work well on different domains like building a strong brain. Existing transferring methods include complex model reconstruction or high-quality retraining on the new domains that makes it hard to implement or ensure high accuracy. This paper introduces a domain-model-based Bayesian network and related solutions to solve this problem. Our solutions make it easier to add new domains while ensure high accuracy like a flexible brain. The experimental results show that our solutions can ensure higher accuracy than the single model one. Furthermore, we also evaluated the network in transferring case and the result shows that the accuracy of our solutions is higher than the single transferred model.
{"title":"Flexible brain: a domain-model based bayesian network for classification","authors":"Guanghao Jin, Qingzeng Song","doi":"10.1080/0952813X.2021.1949753","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1949753","url":null,"abstract":"ABSTRACT Currently, deep learning methods have been widely applied to many fields like classification. Generally, these methods use the technology like transferring to make a model work well on different domains like building a strong brain. Existing transferring methods include complex model reconstruction or high-quality retraining on the new domains that makes it hard to implement or ensure high accuracy. This paper introduces a domain-model-based Bayesian network and related solutions to solve this problem. Our solutions make it easier to add new domains while ensure high accuracy like a flexible brain. The experimental results show that our solutions can ensure higher accuracy than the single model one. Furthermore, we also evaluated the network in transferring case and the result shows that the accuracy of our solutions is higher than the single transferred model.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"34 1","pages":"1011 - 1028"},"PeriodicalIF":2.2,"publicationDate":"2021-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87699867","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 : 2021-07-09DOI: 10.1080/0952813X.2021.1948921
Seyed Reza Shahamiri
ABSTRACT As applications of neural networks increase in our daily lives, their practicality and accuracy become more of a challenge as they are applied to approximate more complicated functions typically composed of different dependent or independent views. While the complexity of the functions and the number of views to be approximated or simulated increases, the task becomes more complicated and more difficult in that it may eventually jeopardise the classifier’s accuracy and make the results unreliable. This paper surveys an improved active learning method called Enhanced Multi-Learner (EML) to facilitate the approximation or simulation of complex functions via neural networks by distributing the complexities of the task under simulation among an array of learners where each network is responsible for learning a specific view. We experimented with EML realisations through neural networks to solve complex problems where traditional methods did not provide adequate results. These experimental studies were conducted in three different domains and are summarised here. Legacy solutions were also provided in each experiment, and the results were compared. The experimental results indicate the superiority of EML base neural networks in dealing with sophisticated pattern recognition problems.
{"title":"Neural network-based multi-view enhanced multi-learner active learning: theory and experiments","authors":"Seyed Reza Shahamiri","doi":"10.1080/0952813X.2021.1948921","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1948921","url":null,"abstract":"ABSTRACT As applications of neural networks increase in our daily lives, their practicality and accuracy become more of a challenge as they are applied to approximate more complicated functions typically composed of different dependent or independent views. While the complexity of the functions and the number of views to be approximated or simulated increases, the task becomes more complicated and more difficult in that it may eventually jeopardise the classifier’s accuracy and make the results unreliable. This paper surveys an improved active learning method called Enhanced Multi-Learner (EML) to facilitate the approximation or simulation of complex functions via neural networks by distributing the complexities of the task under simulation among an array of learners where each network is responsible for learning a specific view. We experimented with EML realisations through neural networks to solve complex problems where traditional methods did not provide adequate results. These experimental studies were conducted in three different domains and are summarised here. Legacy solutions were also provided in each experiment, and the results were compared. The experimental results indicate the superiority of EML base neural networks in dealing with sophisticated pattern recognition problems.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"25 1","pages":"989 - 1009"},"PeriodicalIF":2.2,"publicationDate":"2021-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90798298","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 : 2021-07-06DOI: 10.1080/0952813X.2021.1949754
De-gan Zhang, Peng Yang, Jie Chen, Xiao-dan Zhang, Ting Zhang
ABSTRACT Research on machine deep learning with fuzzy neural network (FNN) is one hot topic in the Artificial Intelligent (AI) domain. In order to support the application of the IoT (Internet of Things) and make use of these image data to get perfect image reasonably and efficiently, it is necessary to fuse these sensed data, therefore the multiple-sensors’ image aggregation becomes a key technology. In this paper, novel FNN-based machine deep learning approach for image aggregation in application of the IoT is proposed. When this approach is done, dynamic learning from eigenvalue transition example can improve traditional learning approach based on static eigenvalue of example. And the neural network is used to be demonstrated its unique superiority of image understanding. FNN-based machine deep learning approach can learn from dynamic eigenvalues, the change of data can be learned and the varieties of the eigenvalue can be understood and remembered. The relative experiments have shown the designed approach for image aggregation is fast and effective, and it can be adapted for the many image applications of the IoT.
{"title":"Novel FNN-based machine deep learning approach for image aggregation in application of the IoT","authors":"De-gan Zhang, Peng Yang, Jie Chen, Xiao-dan Zhang, Ting Zhang","doi":"10.1080/0952813X.2021.1949754","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1949754","url":null,"abstract":"ABSTRACT Research on machine deep learning with fuzzy neural network (FNN) is one hot topic in the Artificial Intelligent (AI) domain. In order to support the application of the IoT (Internet of Things) and make use of these image data to get perfect image reasonably and efficiently, it is necessary to fuse these sensed data, therefore the multiple-sensors’ image aggregation becomes a key technology. In this paper, novel FNN-based machine deep learning approach for image aggregation in application of the IoT is proposed. When this approach is done, dynamic learning from eigenvalue transition example can improve traditional learning approach based on static eigenvalue of example. And the neural network is used to be demonstrated its unique superiority of image understanding. FNN-based machine deep learning approach can learn from dynamic eigenvalues, the change of data can be learned and the varieties of the eigenvalue can be understood and remembered. The relative experiments have shown the designed approach for image aggregation is fast and effective, and it can be adapted for the many image applications of the IoT.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"10 1","pages":"1029 - 1046"},"PeriodicalIF":2.2,"publicationDate":"2021-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88083397","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 : 2021-07-05DOI: 10.1080/0952813X.2021.1948920
Ya-Han Hu, Chih-Fong Tsai
ABSTRACT Online review helpfulness prediction is an important research issue in electronic commerce and data mining. However, the collected datasets used for the analysis and prediction of the helpfulness of online reviews often contain some missing attribute values, such as reviewer background and rating information. In related literatures, many studies have either used the case deletion approach to remove the data containing missing values or considered the imputation of missing values by the mean/mode method. However, none of them consider the direct handling approach without missing value imputation for online review datasets by decision tree-related techniques. Therefore, in this paper, we investigate the suitability of different types of approaches to solve the incomplete dataset problem of online reviews. Specifically, for missing value imputation, several supervised learning techniques including MICE, KNN, SVM, and CART are examined. Moreover, for the direct handling approach without missing value imputation, CART is also performed for this task. The experimental results based on the TripAdvisor dataset for review helpfulness prediction show that the approach where incomplete online review datasets are handled directly without imputation by CART significantly outperforms the other approaches, including case deletion and missing value imputation approaches.
{"title":"An investigation of solutions for handling incomplete online review datasets with missing values","authors":"Ya-Han Hu, Chih-Fong Tsai","doi":"10.1080/0952813X.2021.1948920","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1948920","url":null,"abstract":"ABSTRACT Online review helpfulness prediction is an important research issue in electronic commerce and data mining. However, the collected datasets used for the analysis and prediction of the helpfulness of online reviews often contain some missing attribute values, such as reviewer background and rating information. In related literatures, many studies have either used the case deletion approach to remove the data containing missing values or considered the imputation of missing values by the mean/mode method. However, none of them consider the direct handling approach without missing value imputation for online review datasets by decision tree-related techniques. Therefore, in this paper, we investigate the suitability of different types of approaches to solve the incomplete dataset problem of online reviews. Specifically, for missing value imputation, several supervised learning techniques including MICE, KNN, SVM, and CART are examined. Moreover, for the direct handling approach without missing value imputation, CART is also performed for this task. The experimental results based on the TripAdvisor dataset for review helpfulness prediction show that the approach where incomplete online review datasets are handled directly without imputation by CART significantly outperforms the other approaches, including case deletion and missing value imputation approaches.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"23 1","pages":"971 - 987"},"PeriodicalIF":2.2,"publicationDate":"2021-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80936652","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 : 2021-06-30DOI: 10.1080/0952813X.2021.1938697
L. Chrpa, M. Pilát, Jakub Gemrot
ABSTRACT In dynamic environments, external events might occur and modify the environment without consent of intelligent agents. Plans of the agents might hence be disrupted and, worse, the agents might end up in dead-end states and no longer be able to achieve their goals. Hence, the agents should monitor the environment during plan execution and if they encounter a dangerous situation they should (reactively) act to escape from it. In this paper, we introduce the notion of dangerous states that the agent might encounter during its plan execution in dynamic environments. We present a method for computing lower bound of dangerousness of a state after applying a sequence of actions. That method is leveraged in identifying situations in which the agent has to start acting to avoid danger. We present two types of such behaviour – purely reactive and proactive (eliminating the source of danger). The introduced concepts for planning with dangerous states are implemented and tested in two scenarios – a simple RPG-like game, called Dark Dungeon, and a platform game inspired by the Perestroika video game. The results show that reasoning with dangerous states achieves better success rate (reaching the goals) than naive planning or rule-based techniques.
{"title":"Planning and acting in dynamic environments: identifying and avoiding dangerous situations","authors":"L. Chrpa, M. Pilát, Jakub Gemrot","doi":"10.1080/0952813X.2021.1938697","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1938697","url":null,"abstract":"ABSTRACT In dynamic environments, external events might occur and modify the environment without consent of intelligent agents. Plans of the agents might hence be disrupted and, worse, the agents might end up in dead-end states and no longer be able to achieve their goals. Hence, the agents should monitor the environment during plan execution and if they encounter a dangerous situation they should (reactively) act to escape from it. In this paper, we introduce the notion of dangerous states that the agent might encounter during its plan execution in dynamic environments. We present a method for computing lower bound of dangerousness of a state after applying a sequence of actions. That method is leveraged in identifying situations in which the agent has to start acting to avoid danger. We present two types of such behaviour – purely reactive and proactive (eliminating the source of danger). The introduced concepts for planning with dangerous states are implemented and tested in two scenarios – a simple RPG-like game, called Dark Dungeon, and a platform game inspired by the Perestroika video game. The results show that reasoning with dangerous states achieves better success rate (reaching the goals) than naive planning or rule-based techniques.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"10 1","pages":"925 - 948"},"PeriodicalIF":2.2,"publicationDate":"2021-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86484791","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 : 2021-06-24DOI: 10.1080/0952813X.2021.1938694
Rajole Bhausaheb Namdeo, Gond Vitthal Janardan
ABSTRACT The diagnosis of thyroid via appropriate interpretation of thyroid data is the vital classification issue. Only little contributions are made so far in the automatic diagnosis of thyroid disease. In order to solve Thyroid disorder this paper intends to propose a new thyroid diagnosis model, utilising two-phases includes Feature Extraction and Classification. In the first phase, two sorts of features are extracted that include image features like neighbourhood-based and gradient features, and Principal Component Analysis (PCA) is used to extract the data features as well. Subsequently, two sorts of classification processes are performed. Specifically, Convolutional Neural Network (CNN) is used for image classification by extracting deep features. Neural Network (NN) is used for classifying the disease by obtaining both the image and data features as the input. Finally, both the classified results (CNN and NN) are combined to increase the accuracy rate of diagnosis. Further, as the main aim of this work is to increase the accuracy rate, this paper aims to trigger the optimisation concept. The convolutional layer of CNN is optimally selected, and while classifying under NN the given features should be the optimal one. Hence, the required features are optimally selected. For these optimisations, a new modified algorithm is proposed in this work namely Worst Fitness-based Cuckoo Search (WF-CS) which is the modified form of Cuckoo Search Algorithm (CS). Finally, the performance of proposed WF-CS is compared over other conventional methods like Conventional CS, Genetic Algorithm (GA), FireFly (FF), Artificial Bee Colony (ABC), and Particle Swarm Optimisation (PSO) and proves the superiority of proposed work in detecting the presence of thyroid.
{"title":"Thyroid Disorder Diagnosis by Optimal Convolutional Neuron based CNN Architecture","authors":"Rajole Bhausaheb Namdeo, Gond Vitthal Janardan","doi":"10.1080/0952813X.2021.1938694","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1938694","url":null,"abstract":"ABSTRACT The diagnosis of thyroid via appropriate interpretation of thyroid data is the vital classification issue. Only little contributions are made so far in the automatic diagnosis of thyroid disease. In order to solve Thyroid disorder this paper intends to propose a new thyroid diagnosis model, utilising two-phases includes Feature Extraction and Classification. In the first phase, two sorts of features are extracted that include image features like neighbourhood-based and gradient features, and Principal Component Analysis (PCA) is used to extract the data features as well. Subsequently, two sorts of classification processes are performed. Specifically, Convolutional Neural Network (CNN) is used for image classification by extracting deep features. Neural Network (NN) is used for classifying the disease by obtaining both the image and data features as the input. Finally, both the classified results (CNN and NN) are combined to increase the accuracy rate of diagnosis. Further, as the main aim of this work is to increase the accuracy rate, this paper aims to trigger the optimisation concept. The convolutional layer of CNN is optimally selected, and while classifying under NN the given features should be the optimal one. Hence, the required features are optimally selected. For these optimisations, a new modified algorithm is proposed in this work namely Worst Fitness-based Cuckoo Search (WF-CS) which is the modified form of Cuckoo Search Algorithm (CS). Finally, the performance of proposed WF-CS is compared over other conventional methods like Conventional CS, Genetic Algorithm (GA), FireFly (FF), Artificial Bee Colony (ABC), and Particle Swarm Optimisation (PSO) and proves the superiority of proposed work in detecting the presence of thyroid.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"141 1","pages":"871 - 890"},"PeriodicalIF":2.2,"publicationDate":"2021-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86258155","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 : 2021-06-18DOI: 10.1080/0952813X.2021.1925972
Ángel Díaz-Pacheco, C. García
ABSTRACT Improvement of accuracy in classifiers is a crucial topic in the machine learning field. The problem has been addressed, making new algorithms and selecting the fittest classifier for a given dataset. The latter approach combined with feature selection and pre-processing form up a new paradigm known as Full Model Selection. This paradigm is like a black box whose input is a dataset, and as an output, a precise classification model is obtained. Despite that, full model selection is not the first alternative with the larger datasets of nowadays. We propose the use of MapReduce to deal with huge datasets, a bio-inspired optimisation algorithm and the use of a novel algorithm based on fuzzy classification rules as a proxy model to guide the optimisation process. To the best of our knowledge, this work is the first to propose a classification algorithm based on fuzzy rules as a proxy model. Obtained results showed an accuracy improvement and a considerable reduction of the computing time in datasets of a wide range of sizes.
{"title":"A classification-based fuzzy-rules proxy model to assist in the full model selection problem in high volume datasets","authors":"Ángel Díaz-Pacheco, C. García","doi":"10.1080/0952813X.2021.1925972","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1925972","url":null,"abstract":"ABSTRACT Improvement of accuracy in classifiers is a crucial topic in the machine learning field. The problem has been addressed, making new algorithms and selecting the fittest classifier for a given dataset. The latter approach combined with feature selection and pre-processing form up a new paradigm known as Full Model Selection. This paradigm is like a black box whose input is a dataset, and as an output, a precise classification model is obtained. Despite that, full model selection is not the first alternative with the larger datasets of nowadays. We propose the use of MapReduce to deal with huge datasets, a bio-inspired optimisation algorithm and the use of a novel algorithm based on fuzzy classification rules as a proxy model to guide the optimisation process. To the best of our knowledge, this work is the first to propose a classification algorithm based on fuzzy rules as a proxy model. Obtained results showed an accuracy improvement and a considerable reduction of the computing time in datasets of a wide range of sizes.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"3 1","pages":"815 - 844"},"PeriodicalIF":2.2,"publicationDate":"2021-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72943592","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}
ABSTRACT Nowadays, breast cancer is one of the leading causes of women’s death in the world. If breast cancer is detected at the initial stages, it can ensure long-term survival. Numerous methods have been proposed for the early prediction of such cancer. However, efforts are still ongoing, given the importance of the problem. Artificial Neural Networks (ANN) are a prevalent machine learning algorithm, which is very popular for prediction and classification problems. In this paper, an Intelligent Ensemble Classification method based on Multi-Layer Perceptron neural network (IEC-MLP) is proposed for breast cancer diagnosis. The proposed method consists of two stages: parameters optimisation and ensemble classification. In the first stage, the MLP Neural Network (MLP-NN) parameters, including optimal features, hidden layers, hidden nodes and weights, are optimised with the help of an Evolutionary Algorithm (EA), aiming at maximising the classification accuracy. In the second stage, an ensemble classification algorithm of MLP-NN with optimised parameters is applied to classify the patients. Our proposed IEC-MLP method not only reduces the complexity of MLP-NN and effectively selects the optimal subset of features but also minimises the misclassification cost. The classification results have been evaluated using the IEC-MLP over different breast cancer datasets, and the prediction results have been auspicious (98.74% accuracy on the WBCD dataset). It is noteworthy that the proposed method outperforms the GAANN and CAFS algorithms and other state-of-the-art classifiers. In addition, IEC-MLP is also capable of being employed in diagnosing other cancer types.
{"title":"An intelligent ensemble classification method based on multi-layer perceptron neural network and evolutionary algorithms for breast cancer diagnosis","authors":"Saeed Talatian Azad, Gholamreza Ahmadi, Amin Rezaeipanah","doi":"10.1080/0952813X.2021.1938698","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1938698","url":null,"abstract":"ABSTRACT Nowadays, breast cancer is one of the leading causes of women’s death in the world. If breast cancer is detected at the initial stages, it can ensure long-term survival. Numerous methods have been proposed for the early prediction of such cancer. However, efforts are still ongoing, given the importance of the problem. Artificial Neural Networks (ANN) are a prevalent machine learning algorithm, which is very popular for prediction and classification problems. In this paper, an Intelligent Ensemble Classification method based on Multi-Layer Perceptron neural network (IEC-MLP) is proposed for breast cancer diagnosis. The proposed method consists of two stages: parameters optimisation and ensemble classification. In the first stage, the MLP Neural Network (MLP-NN) parameters, including optimal features, hidden layers, hidden nodes and weights, are optimised with the help of an Evolutionary Algorithm (EA), aiming at maximising the classification accuracy. In the second stage, an ensemble classification algorithm of MLP-NN with optimised parameters is applied to classify the patients. Our proposed IEC-MLP method not only reduces the complexity of MLP-NN and effectively selects the optimal subset of features but also minimises the misclassification cost. The classification results have been evaluated using the IEC-MLP over different breast cancer datasets, and the prediction results have been auspicious (98.74% accuracy on the WBCD dataset). It is noteworthy that the proposed method outperforms the GAANN and CAFS algorithms and other state-of-the-art classifiers. In addition, IEC-MLP is also capable of being employed in diagnosing other cancer types.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"2017 1","pages":"949 - 969"},"PeriodicalIF":2.2,"publicationDate":"2021-06-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74356668","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 : 2021-06-10DOI: 10.1080/0952813X.2021.1938695
Wei-Chao Lin
ABSTRACT Pseudo-relevance feedback (PRF) is a relevance feedback (RF) technique for information retrieval that treats the top k retrieved images as relevance feedback. PRF is used to avoid the limitations of the traditional RF approach, which is a human-in-the-loop process. Although the pseudo-relevance feedback set contains noise, PRF can perform retrieval reasonably effectively. For implementing PRF, the Rocchio algorithm has been considered reasonably effective and is a well-established baseline method. However, it simply treats all of the top k feedback images as being equally similar to the query. Therefore, we present a block-based PRF approach for improving image retrieval performance. In this approach, images in the positive and negative feedback sets are further divided into predefined blocks, each of which contains one to several images, and blocks containing higher- or lower-ranked images will be assigned higher or lower weights, respectively. Experiments using the NUS-WIDE-LITE and Caltech 256 datasets and two different feature representations consistently show that the proposed approach using 30 blocks outperforms the baseline PRF in terms of P@10, P@20, and P@50. Furthermore, we show that a system that incorporates the user’s feedback allows the 30-block-based PRF approach to perform even better.
{"title":"Block-based pseudo-relevance feedback for image retrieval","authors":"Wei-Chao Lin","doi":"10.1080/0952813X.2021.1938695","DOIUrl":"https://doi.org/10.1080/0952813X.2021.1938695","url":null,"abstract":"ABSTRACT Pseudo-relevance feedback (PRF) is a relevance feedback (RF) technique for information retrieval that treats the top k retrieved images as relevance feedback. PRF is used to avoid the limitations of the traditional RF approach, which is a human-in-the-loop process. Although the pseudo-relevance feedback set contains noise, PRF can perform retrieval reasonably effectively. For implementing PRF, the Rocchio algorithm has been considered reasonably effective and is a well-established baseline method. However, it simply treats all of the top k feedback images as being equally similar to the query. Therefore, we present a block-based PRF approach for improving image retrieval performance. In this approach, images in the positive and negative feedback sets are further divided into predefined blocks, each of which contains one to several images, and blocks containing higher- or lower-ranked images will be assigned higher or lower weights, respectively. Experiments using the NUS-WIDE-LITE and Caltech 256 datasets and two different feature representations consistently show that the proposed approach using 30 blocks outperforms the baseline PRF in terms of P@10, P@20, and P@50. Furthermore, we show that a system that incorporates the user’s feedback allows the 30-block-based PRF approach to perform even better.","PeriodicalId":15677,"journal":{"name":"Journal of Experimental & Theoretical Artificial Intelligence","volume":"10 1","pages":"891 - 903"},"PeriodicalIF":2.2,"publicationDate":"2021-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79663601","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：