Showing papers in "Applied Intelligence in 2020"
TL;DR: A study that compared multiple convolutional neural network models to classify CT samples with COVID-19, Influenza viral pneumonia, or no-infection, and achieved an AUC of 0.996 (95%CI: 0.989–1.00) for Coronavirus vs Non-coronav virus cases per thoracic CT studies is technically reviewed.
Abstract: Radiographic patterns on CT chest scans have shown higher sensitivity and specificity compared to RT-PCR detection of COVID-19 which, according to the WHO has a relatively low positive detection rate in the early stages. We technically review a study that compared multiple convolutional neural network (CNN) models to classify CT samples with COVID-19, Influenza viral pneumonia, or no-infection. We compare this mentioned study with one that is developed on existing 2D and 3D deep-learning models, combining them with the latest clinical understanding, and achieved an AUC of 0.996 (95%CI: 0.989–1.00) for Coronavirus vs Non-coronavirus cases per thoracic CT studies. They calculated a sensitivity of 98.2% and a specificity of 92.2%.
424 citations
TL;DR: It was found that most mathematical modeling done were based on the Susceptible-Exposed-Infected-Removed (SEIR) and Susceptibles-infected-recovered (SIR) models while most of the AI implementations were Convolutional Neural Network on X-ray and CT images.
Abstract: In the past few months, several works were published in regards to the dynamics and early detection of COVID-19 via mathematical modeling and Artificial intelligence (AI). The aim of this work is to provide the research community with comprehensive overview of the methods used in these studies as well as a compendium of available open source datasets in regards to COVID-19. In all, 61 journal articles, reports, fact sheets, and websites dealing with COVID-19 were studied and reviewed. It was found that most mathematical modeling done were based on the Susceptible-Exposed-Infected-Removed (SEIR) and Susceptible-infected-recovered (SIR) models while most of the AI implementations were Convolutional Neural Network (CNN) on X-ray and CT images. In terms of available datasets, they include aggregated case reports, medical images, management strategies, healthcare workforce, demography, and mobility during the outbreak. Both Mathematical modeling and AI have both shown to be reliable tools in the fight against this pandemic. Several datasets concerning the COVID-19 have also been collected and shared open source. However, much work is needed to be done in the diversification of the datasets. Other AI and modeling applications in healthcare should be explored in regards to this COVID-19.
198 citations
TL;DR: A generalized Dempster–Shafer evidence theory is proposed, which provides a promising way to model and handle more uncertain information and an algorithm for decision-making is devised based on this theory.
Abstract: Dempster–Shafer evidence theory has been widely used in various fields of applications, because of the flexibility and effectiveness in modeling uncertainties without prior information. However, the existing evidence theory is insufficient to consider the situations where it has no capability to express the fluctuations of data at a given phase of time during their execution, and the uncertainty and imprecision which are inevitably involved in the data occur concurrently with changes to the phase or periodicity of the data. In this paper, therefore, a generalized Dempster–Shafer evidence theory is proposed. To be specific, a mass function in the generalized Dempster–Shafer evidence theory is modeled by a complex number, called as a complex basic belief assignment, which has more powerful ability to express uncertain information. Based on that, a generalized Dempster’s combination rule is exploited. In contrast to the classical Dempster’s combination rule, the condition in terms of the conflict coefficient between the evidences is released in the generalized Dempster’s combination rule. Hence, it is more general and applicable than the classical Dempster’s combination rule. When the complex mass function is degenerated from complex numbers to real numbers, the generalized Dempster’s combination rule degenerates to the classical evidence theory under the condition that the conflict coefficient between the evidences is less than 1. In a word, this generalized Dempster–Shafer evidence theory provides a promising way to model and handle more uncertain information. Thanks to this advantage, an algorithm for decision-making is devised based on the generalized Dempster–Shafer evidence theory. Finally, an application in a medical diagnosis illustrates the efficiency and practicability of the proposed algorithm.
143 citations
TL;DR: A general imbalanced classification model based on deep reinforcement learning, in which the problem is formulated as a sequential decision-making process and solved by a deep Q-learning network, and the agent finally finds an optimal classification policy in imbalanced data.
Abstract: Data in real-world application often exhibit skewed class distribution which poses an intense challenge for machine learning. Conventional classification algorithms are not effective in case of imbalanced data distribution, and may fail when the data distribution is highly imbalanced. To address this issue, we propose a general imbalanced classification model based on deep reinforcement learning, in which we formulate the classification problem as a sequential decision-making process and solve it by a deep Q-learning network. In our model, the agent performs a classification action on one sample in each time step, and the environment evaluates the classification action and returns a reward to the agent. The reward from the minority class sample is larger, so the agent is more sensitive to the minority class. The agent finally finds an optimal classification policy in imbalanced data under the guidance of the specific reward function and beneficial simulated environment. Experiments have shown that our proposed model outperforms other imbalanced classification algorithms, and identifies more minority samples with better classification performance.
115 citations
TL;DR: A novel axiomatic definition of Pythagorean fuzzy distance measurement is initiated, including PFNs and IVPFNs, and the closeness indexes are developed for both expressions, inspired by the idea of technique for order preference by similarity to ideal solution (TOPSIS) approach.
Abstract: Pythagorean fuzzy set, initially extended by Yager from intuitionistic fuzzy set, is capable of modeling information with more uncertainties in the process of multi-criteria decision making (MCDM), thus can be used on wider range of conditions. The fuzzy decision analysis of this paper is mainly built upon two expressions in Pythagorean fuzzy environment, named Pythagorean fuzzy number (PFN) and interval-valued Pythagorean fuzzy number (IVPFN), respectively. We initiate a novel axiomatic definition of Pythagorean fuzzy distance measurement, including PFNs and IVPFNs. After that, corresponding theorems are put forward and then proved. Based on the defined distance measurements, the closeness indexes are developed for both expressions, inspired by the idea of technique for order preference by similarity to ideal solution (TOPSIS) approach. After these basic definitions have been established, the hierarchical decision approach is presented to handle MCDM problems under Pythagorean fuzzy environment. To address hierarchical decision issues, the closeness index-based score function is defined to calculate the score of each permutation for the optimal alternative. To determine criterion weights, a new method based on the proposed similarity measure and aggregation operator of PFNs and IVPFNs is presented according to Pythagorean fuzzy information from decision matrix, rather than being provided in advance by decision makers, which can effectively reduce human subjectivity. An experimental case is then conducted to demonstrate the applicability and flexibility of the proposed decision approach. Finally, extension forms of Pythagorean fuzzy decision approach for heterogeneous information are briefly introduced to show its potentials on further applications in other processing fields with information uncertainties.
105 citations
TL;DR: A new physical-based meta-heuristic optimization algorithm, which is named Transient Search Optimization (TSO) algorithm, inspired by the transient behavior of switched electrical circuits that include storage elements such as inductance and capacitance, is offered.
Abstract: This article offers a new physical-based meta-heuristic optimization algorithm, which is named Transient Search Optimization (TSO) algorithm This algorithm is inspired by the transient behavior of switched electrical circuits that include storage elements such as inductance and capacitance The exploration and exploitation of the TSO algorithm are verified by using twenty-three benchmark, where its statistical (average and standard deviation) results are compared with the most recent 15 optimization algorithms Furthermore, the non-parametric sign test, p value test, execution time, and convergence curves proved the superiority of the TSO against other algorithms Also, the TSO algorithm is applied for the optimal design of three well-known constrained engineering problems (coil spring, welded beam, and pressure vessel) In conclusion, the comparison revealed that the TSO is promising and very competitive algorithm for solving different engineering problems
93 citations
TL;DR: The experimental results demonstrate that the proposed algorithm outperforms other algorithms with low computational efforts while solving economic and micro grid power dispatch problems.
Abstract: This paper proposes a novel hybrid multi-objective algorithm named Multi-objective Spotted Hyena and Emperor Penguin Optimizer (MOSHEPO) for solving both convex and non-convex economic dispatch and micro grid power dispatch problems. The proposed algorithm combines two newly developed bio-inspired optimization algorithms namely Multi-objective Spotted Hyena Optimizer (MOSHO) and Emperor Penguin Optimizer (EPO). MOSHEPO contemplates many non-linear characteristics of power generators such as transmission losses, multiple fuels, valve-point loading, and prohibited operating zones along with their operational constraints, for practical operation. To evaluate the effectiveness of MOSHEPO, the proposed algorithm has been tested on various benchmark test systems and its performance is compared with other well-known approaches. The experimental results demonstrate that the proposed algorithm outperforms other algorithms with low computational efforts while solving economic and micro grid power dispatch problems.
87 citations
TL;DR: A deconvolution region-based convolutional neural network (DR-CNN) to cope with small traffic sign detection and a two-stage adaptive classification loss function for region proposal networks (RPN) and fully connected neural networks within DR-CNN is proposed.
Abstract: Automatic traffic sign detection has great potential for intelligent vehicles. The ability to detect small traffic signs in large traffic scenes enhances the safety of intelligent devices. However, small object detection is a challenging problem in computer vision; the main problem involved in accurate traffic sign detection is the small size of the signs. In this paper, we present a deconvolution region-based convolutional neural network (DR-CNN) to cope with this problem. This method first adds a deconvolution layer and a normalization layer to the output of the convolution layer. It concatenates the features of the different layers into a fused feature map to provide sufficient information for small traffic sign detection. To improve training effectiveness and distinguish hard negative samples from easy positive ones, we propose a two-stage adaptive classification loss function for region proposal networks (RPN) and fully connected neural networks within DR-CNN. Finally, we evaluate our proposed method on the new and challenging Tsinghua-Tencent 100K dataset. We further conduct ablation experiments and analyse the effectiveness of the fused feature map and the two-stage classification loss function. The final experimental results demonstrate the superiority of the proposed method for detecting small traffic signs.
87 citations
TL;DR: Evaluated by level and directional prediction criteria, as well as a newly introduced statistic called the complexity-invariant distance (CID), the VMD-LSTM model shows an outstanding performance in stock price index forecasting.
Abstract: Changes in the composite stock price index are a barometer of social and economic development. To improve the accuracy of stock price index prediction, this paper introduces a new hybrid model, VMD-LSTM, that combines variational mode decomposition (VMD) and a long short-term memory (LSTM) network. The proposed model is based on decomposition-and-ensemble framework. VMD is a data-processing technique through which the original complex series can be decomposed into a limited number of subseries with relatively simple modes of fluctuations. It can effectively overcome the shortcomings of mode mixing that sometimes exist in the empirical mode decomposition (EMD) method. LSTM is an improved version of recurrent neural networks (RNNs) that introduces a “gate” mechanism, and can effectively filter out the critical previous information, making it suitable for the financial time series forecasting. The capability of VMD-LSTM in stock price index forecasting is verified comprehensively by comparing with some single models and the EMD-based and other VMD-based hybrid models. Evaluated by level and directional prediction criteria, as well as a newly introduced statistic called the complexity-invariant distance (CID), the VMD-LSTM model shows an outstanding performance in stock price index forecasting. The hybrid models perform significantly better than the single models, and the forecasting accuracy of the VMD-based models is generally higher than that of the EMD-based models.
75 citations
TL;DR: Experimental results show that compared with the state-of-the-art methods, the proposed intrusion detection approach based on improved deep belief network achieves significant improvement in classification accuracy and FPR.
Abstract: In today’s interconnected society, cyberattacks have become more frequent and sophisticated, and existing intrusion detection systems may not be adequate in the complex cyberthreat landscape. For instance, existing intrusion detection systems may have overfitting, low classification accuracy, and high false positive rate (FPR) when faced with significantly large volume and variety of network data. An intrusion detection approach based on improved deep belief network (DBN) is proposed in this paper to mitigate the above problems, where the dataset is processed by probabilistic mass function (PMF) encoding and Min-Max normalization method to simplify the data preprocessing. Furthermore, a combined sparsity penalty term based on Kullback-Leibler (KL) divergence and non-mean Gaussian distribution is introduced in the likelihood function of the unsupervised training phase of DBN, and sparse constraints retrieve the sparse distribution of the dataset, thus avoiding the problem of feature homogeneity and overfitting. Finally, simulation experiments are performed on the NSL-KDD and UNSW-NB15 public datasets. The proposed method achieves 96.17% and 86.49% accuracy, respectively. Experimental results show that compared with the state-of-the-art methods, the proposed method achieves significant improvement in classification accuracy and FPR.
75 citations
TL;DR: The HSCA modifies the search mechanism of classical SCA by including the leading guidance and hybridizing with simulated quenching algorithm, demonstrating the superiority of the H SCA as compared to other comparative optimization algorithms.
Abstract: The Sine Cosine Algorithm (SCA) is a recently developed efficient metaheuristic algorithm to find the solution of global optimization problems. However, in some circumstances, this algorithm suffers the problem of low exploitation, skipping of true solutions and insufficient balance between exploration and exploitation. Therefore, the present paper aims to alleviate these issues from SCA by proposing an improved variant of SCA called HSCA. The HSCA modifies the search mechanism of classical SCA by including the leading guidance and hybridizing with simulated quenching algorithm. The proposed HSCA is tested on classical benchmark set, standard and complex benchmarks sets IEEE CEC 2014 and CEC 2017 and four engineering optimization problems. In addition to these problems, the HSCA is also used to train multilayer perceptrons as a real-life application. The experimental results and analysis on benchmark problems and real-life application problems demonstrate the superiority of the HSCA as compared to other comparative optimization algorithms.
TL;DR: This research work aims at presenting a bi-level genetic algorithm approach of an optimized data analytic AI technique for monitoring the health of the agriculture vehicles which can be economically utilized on smartphone end-devices using the built-in microphones instead of expensive IoT sensors.
Abstract: In the era of Internet of things (IoT), network Connection of an enormous number of agriculture machines and service centers is an expectation. However, it will be with a generation of massive volume of data, thus overwhelming the network traffic and storage system especially when manufacturers give maintenance service typically by various data analytic applications on the cloud. The situation is more complex in the context of low latency applications such as health monitoring of agriculture machines, although require emergency responses. Performing the computational intelligence on edge devices is one of the best approaches in developing green communications and managing the blast of network traffic. Due to the increasing usage of smartphone applications, the edge computation on the smartphone can highly assist the network traffic management. In connection with the mentioned point, in the context of exploiting the limited computation power of smartphones, the design of an AI-based data analytic technique is a challenging task. On the other hand, the users’ need for economic technology makes it not to be easily pierced. This research work aims both targets by presenting a bi-level genetic algorithm approach of an optimized data analytic AI technique for monitoring the health of the agriculture vehicles which can be economically utilized on smartphone end-devices using the built-in microphones instead of expensive IoT sensors.
TL;DR: This research presents a novel aggregating method for constructing an aggregated topic model that is composed of the topics with greater coherence than individual models that outperforms those topic models at a statistically significant level in terms of topic coherence over an external corpus.
Abstract: This research presents a novel aggregating method for constructing an aggregated topic model that is composed of the topics with greater coherence than individual models. When generating a topic model, a number of parameters have to be specified. The resulting topics can be very general or very specific, which depend on the chosen parameters. In this study we investigate the process of aggregating multiple topic models generated using different parameters with a focus on whether combining the general and specific topics is able to increase topic coherence. We employ cosine similarity and Jensen-Shannon divergence to compute the similarity among topics and combine them into an aggregated model when their similarity scores exceed a predefined threshold. The model is evaluated against the standard topics models generated by the latent Dirichlet allocation and Non-negative Matrix Factorisation. Specifically we use the coherence of topics to compare the individual models that create aggregated models against those of the aggregated model and models generated by Non-negative Matrix Factorisation, respectively. The results demonstrate that the aggregated model outperforms those topic models at a statistically significant level in terms of topic coherence over an external corpus. We also make use of the aggregated topic model on social media data to validate the method in a realistic scenario and find that again it outperforms individual topic models.
TL;DR: A driver action recognition model is proposed, which is called deformable and dilated Faster R-CNN (DD-RCNN), which utilizes the detection of motion-specific objects to classify driver actions exhibiting great intra-class differences and inter-class similarity.
Abstract: Distracted driver action is the main cause of road traffic crashes, which threatens the security of human life and public property. Based on the observation that cues (like the hand holding the cigarette) reveal what the driver is doing, a driver action recognition model is proposed, which is called deformable and dilated Faster R-CNN (DD-RCNN). Our approach utilizes the detection of motion-specific objects to classify driver actions exhibiting great intra-class differences and inter-class similarity. Firstly, deformable and dilated residual block are designed to extract features of action-specific RoIs that are small in size and irregular in shape (such as cigarettes and cell phones). Attention modules are embedded in the modified ResNet to reweight features in channel and spatial dimensions. Then, the region proposal optimization network (RPON) is presented to reduce the number of RoIs entering R-CNN and improves model efficiency. Lastly, the RoI pooling module is replaced with the deformable one, and the simplified R-CNN without regression layer is trained as the final classifier. Experiments show that DD-RCNN demonstrates state-of-the-art results on Kaggle-driving dataset and self-built dataset.
TL;DR: A lightweight green mangoes detection method based on YOLOv3 is proposed here, which has the best detection results in terms of dense, backlit, direct light, night, long distance, and special angle scenes under complex lighting.
Abstract: When a robot picks green fruit under natural light, the color of the fruit is similar to the background; uneven lighting and fruit and leaf occlusion often affect the performance of the detection method. We take green mangoes as an experimental object. A lightweight green mangoes detection method based on YOLOv3 is proposed here. To improve the detection speed of the method, we first combine the color, texture, and shape features of green mango to design a lightweight network unit to replace the residual units in YOLOv3. Second, the improved Multiscale context aggregation (MSCA) module is used to concatenate multilayer features and make predictions, solving the problem of insufficient position information and semantic information on the prediction feature map in YOLOv3; this approach effectively improves the detection effect for the green mangoes. To address the overlap of green mangoes, soft non-maximum suppression (Soft-NMS) is used to replace non-maximum suppression (NMS), thereby reducing the missing of predicted boxes due to green mango overlaps. Finally, an auxiliary inspection green mango image enhancement algorithm (CLAHE-Mango) is proposed, is suitable for low-brightness detection environments and improves the accuracy of the green mango detection method. The experimental results show that the F1% of Light-YOLOv3 in the test set is 97.7%. To verify the performance of Light-YOLOv3 under the embedded platform, we embed one-stage methods into the Adreno 640 and Mali-G76 platforms. Compared with YOLOv3, the F1% of Light-YOLOv3 is increased by 4.5%, and the running speed is increased by 5 times, which can meet the real-time running requirements for picking robots. Through three sets of comparative experiments, we could determine that our method has the best detection results in terms of dense, backlit, direct light, night, long distance, and special angle scenes under complex lighting.
TL;DR: A novel multi-label text classification method that combines dynamic semantic representation model and deep neural network (DSRM-DNN) is proposed that outperforms the state-of-the-art methods.
Abstract: The increment of new words and text categories requires more accurate and robust classification methods. In this paper, we propose a novel multi-label text classification method that combines dynamic semantic representation model and deep neural network (DSRM-DNN). DSRM-DNN first utilizes word embedding model and clustering algorithm to select semantic words. Then the selected words are designated as the elements of DSRM-DNN and quantified by the weighted combination of word attributes. Finally, we construct a text classifier by combining deep belief network and back-propagation neural network. During the classification process, the low-frequency words and new words are re-expressed by the existing semantic words under sparse constraint. We evaluate the performance of DSRM-DNN on RCV1-v2, Reuters-21578, EUR-Lex, and Bookmarks. Experimental results show that our method outperforms the state-of-the-art methods.
TL;DR: Evidence is given that if a deep neural network is trained to construct the deep features of the data, imputation based on deep features is better than that directly on the original data and when data encounters larger missing ratio and various missing patterns, the proposed algorithm has the ability to achieve more accurate and stable imputation performance.
Abstract: Due to cluster instability, not in the cluster monitoring system. This paper focuses on the missing data imputation processing for the cluster monitoring application and proposes a new hybrid multiple imputation framework. This new imputation approach is different from the conventional multiple imputation technologies in the fact that it attempts to impute the missing data for an arbitrary missing pattern with a model-based and data-driven combination architecture. Essentially, the deep neural network, as the data model, extracts deep features from the data and deep features are further calculated then by a regression or data-driven strategies and used to create the estimation of missing data with the arbitrary missing pattern. This paper gives evidence that if we can train a deep neural network to construct the deep features of the data, imputation based on deep features is better than that directly on the original data. In the experiments, we compare the proposed method with other conventional multiple imputation approaches for varying missing data patterns, missing ratios, and different datasets including real cluster data. The result illustrates that when data encounters larger missing ratio and various missing patterns, the proposed algorithm has the ability to achieve more accurate and stable imputation performance.
TL;DR: A novel hybrid forecasting algorithm based on locally weighted support vector regression (LWSVR) and the modified grasshopper optimization algorithm (MGOA) to solve the short term load forecasting (STLF) problem.
Abstract: Many day-to-day operation decisions in a smart city need short term load forecasting (STLF) of its customers. STLF is a challenging task because the forecasting accuracy is affected by external factors whose relationships are usually complex and nonlinear. In this paper, a novel hybrid forecasting algorithm is proposed. The proposed hybrid forecasting method is based on locally weighted support vector regression (LWSVR) and the modified grasshopper optimization algorithm (MGOA). Obtaining the appropriate values of LWSVR parameters is vital to achieving satisfactory forecasting accuracy. Therefore, the MGOA is proposed in this paper to optimally select the LWSVR’s parameters. The proposed MGOA can be derived by presenting two modifications on the conventional GOA in which the chaotic initialization and the sigmoid decreasing criterion are employed to treat the drawbacks of the conventional GOA. Then the hybrid LWSVR-MGOA method is used to solve the STLF problem. The performance of the proposed LWSVR-MGOA method is assessed using six different real-world datasets. The results reveal that the proposed forecasting method gives a much better forecasting performance in comparison with some published forecasting methods in all cases.
TL;DR: An asymmetric depthwise separable convolution network (ADSCNet) which is a lightweight neural network for real-time semantic segmentation and Dense Dilated Convolution Connections (DDCC), which connects a set of dilated convolutional layers in a dense way, is introduced in the network.
Abstract: Semantic segmentation can be considered as a per-pixel localization and classification problem, which gives a meaningful label to each pixel in an input image. Deep convolutional neural networks have made extremely successful in semantic segmentation in recent years. However, some challenges still exist. The first challenge task is that most current networks are complex and it is hard to deploy these models on mobile devices because of the limitation of computational cost and memory. Getting more contextual information from downsampled feature maps is another challenging task. To this end, we propose an asymmetric depthwise separable convolution network (ADSCNet) which is a lightweight neural network for real-time semantic segmentation. To facilitating information propagation, Dense Dilated Convolution Connections (DDCC), which connects a set of dilated convolutional layers in a dense way, is introduced in the network. Pooling operation is inserted before ADSCNet unit to cover more contextual information in prediction. Extensive experimental results validate the superior performance of our proposed method compared with other network architectures. Our approach achieves mean intersection over union (mIOU) of 67.5% on Cityscapes dataset at 76.9 frames per second.
TL;DR: Experimental results on both real and artificial networks show that the proposed algorithm can uncover communities more accurately than all the comparison algorithms.
Abstract: Community structure is an important characteristic of complex networks. Uncovering communities in complex networks is currently a hot research topic in the field of network analysis. Local community detection algorithms based on seed-extension are widely used for addressing this problem because they excel in efficiency and effectiveness. Compared with global community detection methods, local methods can uncover communities without the integral structural information of complex networks. However, they still have quality and stability deficiencies in overlapping community detection. For this reason, a local community detection algorithm based on internal force between nodes is proposed. First, local degree central nodes and Jaccard coefficient are used to detect core members of communities as seeds in the network, thus guaranteeing that the selected seeds are central nodes of communities. Second, the node with maximum degree among seeds is pre-extended by the fitness function every time. Finally, the top k nodes with the best performance in pre-extension process are extended by the fitness function with internal force between nodes to obtain high-quality communities in the network. Experimental results on both real and artificial networks show that the proposed algorithm can uncover communities more accurately than all the comparison algorithms.
TL;DR: The experimental results demonstrated that the proposed SOA algorithm is able to solve challenging constrained optimization problems and outperforms the other state-of-the-art optimization algorithms.
Abstract: This paper presents a novel bio-inspired algorithm called Sandpiper Optimization Algorithm (SOA) and applies it to solve challenging real-life problems. The main inspiration behind this algorithm is the migration and attacking behaviour of sandpipers. These two steps are modeled and implemented computationally to emphasize intensification and diversification in the search space. The comparison of proposed SOA algorithm is performed with nine competing optimization algorithms over 44 benchmark functions. The analysis of computational complexity and convergence behaviors of the proposed algorithm have been evaluated. Further, SOA algorithm is hybridized with decision tree machine-learning algorithm to solve real-life applications. The experimental results demonstrated that the proposed algorithm is able to solve challenging constrained optimization problems and outperforms the other state-of-the-art optimization algorithms.
TL;DR: A Dirichlet process biterm-based mixture model (DP-BMM) is proposed, which can deal with the topic drift problem and the sparsity problem in short text stream clustering and can achieve a better performance than the state-of-the-art methods in terms of NMI metrics.
Abstract: Short text stream clustering has become an important problem for mining textual data in diverse social media platforms (e.g., Twitter). However, most of the existing clustering methods (e.g., LDA and PLSA) are developed based on the assumption of a static corpus of long texts, while little attention has been given to short text streams. Different from the long texts, the clustering of short texts is more challenging since their word co-occurrence pattern easily suffers from a sparsity problem. In this paper, we propose a Dirichlet process biterm-based mixture model (DP-BMM), which can deal with the topic drift problem and the sparsity problem in short text stream clustering. The major advantages of DP-BMM include (1) DP-BMM explicitly exploits the word-pairs constructed from each document to enhance the word co-occurrence pattern in short texts; (2) DP-BMM can deal with the topic drift problem of short text streams naturally. Moreover, we further propose an improved algorithm of DP-BMM with forgetting property called DP-BMM-FP, which can efficiently delete biterms of outdated documents by deleting clusters of outdated batches. To perform inference, we adopt an online Gibbs sampling method for parameter estimation. Our extensive experimental results on real-world datasets show that DP-BMM and DP-BMM-FP can achieve a better performance than the state-of-the-art methods in terms of NMI metrics.
TL;DR: A novel game-based ACO that consists of two ACOs: Ant Colony System (ACS) and Max-Min Ant System (MMAS) and an entropy-weighted learning strategy is proposed, which has well performance in terms of both the solution precision and the astringency.
Abstract: Ant Colony Optimization (ACO) algorithms tend to fall into local optimal and have insufficient astringency when applied to solve Traveling Salesman Problem (TSP). To address this issue, a novel game-based ACO (NACO) is proposed in this report. NACO consists of two ACOs: Ant Colony System (ACS) and Max-Min Ant System (MMAS). First, an entropy-weighted learning strategy is proposed. By improving diversity adaptively, the optimal solution precision can be optimized. Then, to improve the astringency, a nucleolus game strategy is set for ACS colonies. ACS colonies under cooperation share pheromone distribution and distribute cooperative profits through nucleolus. Finally, to jump out of the local optimum, mean filtering is introduced to process the pheromone distribution when the algorithm stalls. From the experimental results, it is demonstrated that NACO has well performance in terms of both the solution precision and the astringency.
TL;DR: This paper investigates the distributed assembly blocking flow-shop scheduling problem (DABFSP), which consists of two stages: production and assembly, and proposes a constructive heuristic and iterated local search that can solve the DABFSP effectively and efficiently.
Abstract: Scheduling in distributed production system has become an active research field in recent years. This paper investigates the distributed assembly blocking flow-shop scheduling problem (DABFSP), which consists of two stages: production and assembly. The first stage is processing jobs in several identical factories. Each factory has a series of machines no intermediate buffers existing between adjacent ones. The second stage assembles the processed jobs into the final products through a single machine. The objective is to minimize the maximum completion time or makespan of all products. To address this problem, a constructive heuristic is proposed based on a new assignment rule of jobs and a product-based insertion procedure. Afterwards, an iterated local search (ILS) is presented, which integrates an integrated encoding scheme, a multi-type perturbation procedure containing four kinds of perturbed operators based on problem-specific knowledge and a critical-job-based variable neighborhood search. Finally, a comprehensive computational experiment and comparisons with the closely related and well performing methods in the literature are carried out. The experimental and comparison results show that the proposed constructive heuristic and ILS can solve the DABFSP effectively and efficiently.
TL;DR: The inverted residuals technique is used to improve the convolutional layer of YOLOv3 to reduce the missing of predicted boxes due to vehicle overlaps, and soft non maximum suppression is used in order to cope with the overlapping of vehicles in traffic videos.
Abstract: According to the problem that the multi-scale vehicle objects in traffic surveillance video are difficult to detect and the overlapping objects are prone to missed detection, an improved vehicle object detection method based on YOLOv3 was proposed. In order to extract feature more efficiently, we first use the inverted residuals technique to improve the convolutional layer of YOLOv3. To solve the multi-scale vehicle object detection problem, three spatial pyramid pooling(SPP) modules are added before each YOLO layer to obtain multi-scale information. In order to cope with the overlapping of vehicles in traffic videos, soft non maximum suppression (Soft-NMS) is used to replace non maximum suppression (NMS), thereby reducing the missing of predicted boxes due to vehicle overlaps. Our experiment results in the Car dataset and the KITTI dataset confirm that the proposed method achieves good detection results for vehicle objects of various scales in various scenes. Our method can meet the needs of practical applications better.
TL;DR: In this article, a multi-split approach based on Gini index and p-value was adopted to predict students' academic performance at two stages of course delivery (20% and 50% respectively).
Abstract: Predicting students’ academic performance has been a research area of interest in recent years, with many institutions focusing on improving the students’ performance and the education quality. The analysis and prediction of students’ performance can be achieved using various data mining techniques. Moreover, such techniques allow instructors to determine possible factors that may affect the students’ final marks. To that end, this work analyzes two different undergraduate datasets at two different universities. Furthermore, this work aims to predict the students’ performance at two stages of course delivery (20% and 50% respectively). This analysis allows for properly choosing the appropriate machine learning algorithms to use as well as optimize the algorithms’ parameters. Furthermore, this work adopts a systematic multi-split approach based on Gini index and p-value. This is done by optimizing a suitable bagging ensemble learner that is built from any combination of six potential base machine learning algorithms. It is shown through experimental results that the posited bagging ensemble models achieve high accuracy for the target group for both datasets.
TL;DR: A novel discrete whale optimization algorithm (DWOA) which uses the new proposed V -shaped function to generate an integer vector and can be used to solve discrete optimization problems with solution space.
Abstract: Whale optimization algorithm (WOA) is a recently proposed meta-heuristic algorithm which imitates the hunting behavior of humpback whales. Due to its characteristic advantages, it has found its place in the mature population-based methods in many scientific and engineering fields. Because WOA was proposed for continuous optimization, it cannot be directly used to solve discrete optimization problems. For this purpose, we first give a new V -shaped function by drawing lesson from the existing discretization methods, which transfer a real vector to an integer vector. On this basis, we propose a novel discrete whale optimization algorithm (DWOA). DWOA uses the new proposed V -shaped function to generate an integer vector, and it can be used to solve discrete optimization problems with solution space {0,1,…,m1}×{0,1,…,m2}×… ×{0,1,…,mn}. To verify effectiveness of DWOA for the 0-1 knapsack problem and the discount {0-1} knapsack problem, we solve their benchmark instances from published literature and compare with the state-of-the-art algorithms. The comparison results show that the DWOA has more superiority than existing algorithms for the two kinds of knapsack problems.
TL;DR: The objective of the proposed work is to improve the accuracy by using the most similar neighbor for each target item so as to predict the target item, since finding k for different target item is computationally expensive.
Abstract: The item-based collaborative filtering technique recommends an item to the user from the rating of k-nearest items. Generally, a random value of k is considered to find nearest neighbor from item-item similarity matrix. However, consideration of a random value for k intuitively is not a rational approach, as different items may have different value of k nearest neighbor. Sparsity in the data set is another challenge in collaborative filtering, as number of co-rated items’ may be few or zero. Due to the above two reasons, collaborative filtering provides inaccurate recommendations, because the predicted rating may tend towards the Mean. The objective of the proposed work is to improve the accuracy by mitigating the above issues. Instead of using a random value of k, we use the most similar neighbor for each target item so as to predict the target item, since finding k for different target item is computationally expensive. Bhattacharyya Coefficient is used as a similarity measure to handle sparsity in the dataset. The performance of the proposed algorithm is tested the datasets of MovieLens and Film Trust, and experimental results reveal better prediction accuracy than the best of the prevalent prediction approaches exist in literature.
TL;DR: A novel self-organizing fuzzy neural network with an adaptive learning algorithm (SOFNN-ALA) for nonlinear system modeling and identification in industrial processes is proposed and exhibits a better comprehensive performance than some other state-of-the-art SOFNNs for non linear system modeling in industrial applications.
Abstract: In this paper, a novel self-organizing fuzzy neural network with an adaptive learning algorithm (SOFNN-ALA) for nonlinear system modeling and identification in industrial processes is proposed. To efficiently enhance the generalization capability, the proposed SOFNN-ALA is designed by using both structure identification and parameter estimation simultaneously in the learning process. In the structure identification phase, the rule neuron with the highest neuronal activity will be split into two new rule neurons. Meanwhile, the redundant rule neurons with small singular values will be removed to simplify the network structure. In the parameter estimation phase, an adaptive learning algorithm (ALA), which is designed based on the widely used Levenberg-Marquardt (LM) optimization algorithm, is adopted to optimize the network parameters. The ALA-based learning algorithm can not only speed up the convergence speed but also enhance the modeling performance. Moreover, we carefully analyze the convergence of the proposed SOFNN-ALA to guarantee its successful practical application. Finally, the effectiveness and efficiency of the proposed SOFNN-ALA is validated by several examples. The experimental results demonstrate that the proposed SOFNN-ALA exhibits a better comprehensive performance than some other state-of-the-art SOFNNs for nonlinear system modeling in industrial applications. The source code can be downloaded from https://github.com/hyitzhb/SOFNN-ALA.git.
TL;DR: A new feature redundancy term that considers the relevancy between a candidate feature and the class given each already-selected feature, and a novel feature selection method named min-redundancy and max-dependency (MRMD) is proposed that achieves the best classification performance with respect to multiple evaluation criteria.
Abstract: Feature selection plays a critical role in many applications that are relevant to machine learning, image processing and gene expression analysis. Traditional feature selection methods intend to maximize feature dependency while minimizing feature redundancy. In previous information-theoretical-based feature selection methods, feature redundancy term is measured by the mutual information between a candidate feature and each already-selected feature or the interaction information among a candidate feature, each already-selected feature and the class. However, the larger values of the traditional feature redundancy term do not indicate the worse a candidate feature because a candidate feature can obtain large redundant information, meanwhile offering large new classification information. To address this issue, we design a new feature redundancy term that considers the relevancy between a candidate feature and the class given each already-selected feature, and a novel feature selection method named min-redundancy and max-dependency (MRMD) is proposed. To verify the effectiveness of our method, MRMD is compared to eight competitive methods on an artificial example and fifteen real-world data sets respectively. The experimental results show that our method achieves the best classification performance with respect to multiple evaluation criteria.