How to fast, accurately and robustly recognize wheat diseases, particularly for those diseases with mild-to-moderate severity, is a challenge for prevention and control of crop disease timely. In this study, image pro...
详细信息
How to fast, accurately and robustly recognize wheat diseases, particularly for those diseases with mild-to-moderate severity, is a challenge for prevention and control of crop disease timely. In this study, image processing technique was applied to segment the infected regions of disease leaves. Twenty disease features were extracted, and eighteen larger weight features were selected by Relief-F algorithm to generate the models of Support Vector Machine (SVM), Relevance Vector Machine (RVM) and Back Propagation Neural Network (BPNN). Subsequently, these models were used to identify two kinds of wheat diseases, namely, wheat stripe rust and powdery mildew. Total 136 samples, including 68 training samples and 68 test samples with different infection severities were used to study the recognition capabilities of the three models. Results showed that high predictive accuracies in identification of two wheat diseases with varying severity for all three models. Overall accuracy of RVM was 89.71%, which was superior to 83.82% of SVM and inferior to 92.64% of BPNN. Meanwhile, the recognition accuracies of SVM, RVM and BPNN models for mild-to-moderate disease were 83.33%, 88.33% and 91.67%, respectively. The prediction time of RVM was less than those of SVM and BPNN, with differences as large as 7.96 and 31.68 times, respectively. Therefore, RVM appeared to be the most suitable for real-time identifying wheat leaf diseases among the three models, which can provide important technical support for wheat diseases management.
Attribute reduction is one of the key issues for data preprocess in data mining. Many heuristic attribute reduction algorithms based on discernibility matrix have been proposed for inconsistent decision tables. Howeve...
详细信息
ISBN:
(纸本)9781509003914
Attribute reduction is one of the key issues for data preprocess in data mining. Many heuristic attribute reduction algorithms based on discernibility matrix have been proposed for inconsistent decision tables. However, these methods are usually computationally time-consuming. To address this issue, the derived consistent decision tables are defined for different definitions of relative reducts. The computations for different reducts of the original inconsistent decision tables are converted into the computations for their corresponding reducts of the derived consistent datasets. The relative discernibility object pair and the more optimal relative discernibility degree from view of the boundary region are designed to accelerate the attribute reduction process. An efficient attribute reduction framework using relative discernibility degree is proposed for large datasets. Experimental results show that our attribute reduction algorithms are effective and feasible for large inconsistent datasets.
As an important characteristic of social media (i.e. Flickr or Facebook), user communities or groups are beginning to attract increasing attention. Most of the previous studies on group recommendation only consider a ...
详细信息
In this paper, a new method of non contact and on-line measurement is presented for the change of the axis of the rotation axis. Under the condition of a constant light source, the image of the original position and t...
详细信息
No reference video quality assessment (NR-VQA) measures distorted videos quantitatively without the reference of original high quality videos. Conventional NR-VQA methods are generally designed for specific types of d...
详细信息
With the rapid growth of data volume, knowledge acquisition for bigdata has become a new challenge. To address this issue, the hierarchical decision table is defined and implemented in this work. The properties of di...
详细信息
ISBN:
(纸本)9781467372220
With the rapid growth of data volume, knowledge acquisition for bigdata has become a new challenge. To address this issue, the hierarchical decision table is defined and implemented in this work. The properties of different hierarchical decision tables are discussed under the different granularity of conditional attributes. A novel knowledge acquisition algorithm for bigdata using MapReduce is proposed. Experimental results demonstrate that the proposed algorithm is able to deal with bigdata and mine hierarchical decision rules under the different granularity.
Adding to societal changes today, are the miscellaneous bigdata produced in different fields. Coupled with these data is the appearance of risk management. Admittedly, to predict future trend by using these data is c...
详细信息
Adding to societal changes today, are the miscellaneous bigdata produced in different fields. Coupled with these data is the appearance of risk management. Admittedly, to predict future trend by using these data is conducive to make everything more efficient and easy. Now, no matter companies or individuals, they increasingly focus on identifying risks and managing them before risks. Effective risk management will lead them to deal with potential problems. This thesis focuses on risk management of flight delay area using big real time data. It proposes two different prediction models, one is called General Long Term Departure Prediction Model and the other is named as Improved Real Time Arrival Prediction Model. By studying the main factors lead to flight delay, this thesis takes weather, carrier, National Aviation System, security and previous late aircraft as analysis factors. By utilizing our models can do not only long time but also short term flight delay predictions. The results demonstrate goodness of fit. Besides the theory part, it also presents a practical and beautiful web application for real time flight arrival prediction based on our second model.
Distributed Collaborative Machine Learning (DCML) has emerged in artificial intelligence-empowered edge computing environments, such as the Industrial Internet of Things (IIoT), to process tremendous data generated by...
详细信息
Distributed Collaborative Machine Learning (DCML) has emerged in artificial intelligence-empowered edge computing environments, such as the Industrial Internet of Things (IIoT), to process tremendous data generated by smart devices. However, parallel DCML frameworks require resource-constrained devices to update the entire Deep Neural Network (DNN) models and are vulnerable to reconstruction attacks. Concurrently, the serial DCML frameworks suffer from training efficiency problems due to their serial training nature. In this paper, we propose a Model Pruning-enabled Federated Split Learning framework (MP-FSL) to reduce resource consumption with a secure and efficient training scheme. Specifically, MP-FSL compresses DNN models by adaptive channel pruning and splits each compressed model into two parts that are assigned to the client and the server. Meanwhile, MP-FSL adopts a novel aggregation algorithm to aggregate the pruned heterogeneous models. We implement MP-FSL with a real FL platform to evaluate its performance. The experimental results show that MP-FSL outperforms the state-of-the-art frameworks in model accuracy by up to 1.35%, while concurrently reducing storage and computational resource consumption by up to 32.2% and 26.73%, respectively. These results demonstrate that MP-FSL is a comprehensive solution to the challenges faced by DCML, with superior performance in both reduced resource consumption and enhanced model performance.
暂无评论