Peer-to-peer (P2P) lending offers an alternative way to access credit. Unlike established lending institutions with proven credit risk management practices, P2P platforms rely on numerous independent variables to eval...Peer-to-peer (P2P) lending offers an alternative way to access credit. Unlike established lending institutions with proven credit risk management practices, P2P platforms rely on numerous independent variables to evaluate loan applicants’ creditworthiness. This study aims to estimate default probabilities using a mixture-of-experts neural network in P2P lending. The approach involves coupling unsupervised clustering to capture essential data properties with a classification algorithm based on the mixture-of-experts structure. This classic design enhances model capacity without significant computational overhead. The model was tested using P2P data from Lending Club, comparing it to other methods like Logistic Regression, AdaBoost, Gradient Boosting, Decision Tree, Support Vector Machine, and Random Forest. The hybrid model demonstrated superior performance, with a Mean Squared Error reduction of at least 25%.展开更多
Wireless sensor networks are widely used for its flexibility, but they also suffer from problems like limited capacity, large node number and vulnerability to security threats. In this paper, we propose a multi-path r...Wireless sensor networks are widely used for its flexibility, but they also suffer from problems like limited capacity, large node number and vulnerability to security threats. In this paper, we propose a multi-path routing protocol based on the credible cluster heads. The protocol chooses nodes with more energy remained as cluster heads at the cluster head choosing phase, and then authenticates them by the neighbor cluster heads. Using trust mechanisms it creates the credit value, and based on the credit value the multi-path cluster head routing can finally be found. The credit value is created and exchanged among the cluster heads only. Theoretical analysis combined with simulation results demonstrate that this protocol can save the resource, prolong the lifetime, and ensure the security and performance of the network.展开更多
Traditional clustering method is easy to slow convergence speed because of high data dimension and setting random initial clustering center. To improve these problems, a novel method combining subtractive clustering w...Traditional clustering method is easy to slow convergence speed because of high data dimension and setting random initial clustering center. To improve these problems, a novel method combining subtractive clustering with fuzzy C-means( FCM)clustering will be advanced. In the method, the initial cluster number and cluster center can be obtained using subtractive clustering. On this basis,clustering result will be further optimized with FCM. In addition,the data dimension will be reduced through the analytic hierarchy process( AHP) before clustering calculating.In order to verify the effectiveness of fusion algorithm,an example about enterprise credit evaluation will be carried out. The results show that the fusion clustering algorithm is suitable for classifying high-dimension data,and the algorithm also does well in running up processing speed and improving visibility of result. So the method is suitable to promote the use.展开更多
In this paper we applied the technique of Self Organizing Map (SOM) to segment individuals based on their credit information. SOM is an unsupervised machine learning method that reduces data complexity and dimensional...In this paper we applied the technique of Self Organizing Map (SOM) to segment individuals based on their credit information. SOM is an unsupervised machine learning method that reduces data complexity and dimensionality while keeping sits original topology, which is superior to other dimension reduction methods especially when features in data have unclear nonlinear relations. Through this method we provide more clear and intuitive segmentation that other traditional methods cannot achieve.展开更多
众所周知,共识机制是区块链的核心,是区块链实现分布式存储的关键。随着各种区块链共识机制地出现,基于共识机制的优化方法也相继被提出,主要从优化共识过程以及控制共识节点的数量入手,解决共识机制吞吐量低、高时延、高资源等问题。然...众所周知,共识机制是区块链的核心,是区块链实现分布式存储的关键。随着各种区块链共识机制地出现,基于共识机制的优化方法也相继被提出,主要从优化共识过程以及控制共识节点的数量入手,解决共识机制吞吐量低、高时延、高资源等问题。然而,许多基于共识机制的优化缺乏理论的分析,也没有提及关键参数会影响共识机制的性能。为此,文中将以实用拜占庭算法(Practical Byzantine Fault Tolerance Algorithm,PBFT)、基于分组的实用拜占庭算法(Practical Byzantine Fault Tolerant Algorithm Based on Group,G-PBFT)以及基于分组和信誉的实用拜占庭算法(Practical Byzantine Fault Tolerant Algorithm Based on Clustering and Reputation,GR-PBFT)为例,构建三者的数学模型,进行性能分析。根据交易吞吐量、交易失败概率、区块认证失败概率和通信复杂度等性能指标进行对比。仿真结果表明:在同等节点数量下,G-PBFTD、GR-PBFT算法的吞吐量为PBFT的1.57倍、2.38倍;G-PBFTD、GR-PBFT算法的交易认证失败概率比PBFT下降了16%、39%;G-PBFTD、GR-PBFT算法的通信复杂度比PBFT下降了3.1倍、4.0倍,优化效果显著。展开更多
文摘Peer-to-peer (P2P) lending offers an alternative way to access credit. Unlike established lending institutions with proven credit risk management practices, P2P platforms rely on numerous independent variables to evaluate loan applicants’ creditworthiness. This study aims to estimate default probabilities using a mixture-of-experts neural network in P2P lending. The approach involves coupling unsupervised clustering to capture essential data properties with a classification algorithm based on the mixture-of-experts structure. This classic design enhances model capacity without significant computational overhead. The model was tested using P2P data from Lending Club, comparing it to other methods like Logistic Regression, AdaBoost, Gradient Boosting, Decision Tree, Support Vector Machine, and Random Forest. The hybrid model demonstrated superior performance, with a Mean Squared Error reduction of at least 25%.
文摘Wireless sensor networks are widely used for its flexibility, but they also suffer from problems like limited capacity, large node number and vulnerability to security threats. In this paper, we propose a multi-path routing protocol based on the credible cluster heads. The protocol chooses nodes with more energy remained as cluster heads at the cluster head choosing phase, and then authenticates them by the neighbor cluster heads. Using trust mechanisms it creates the credit value, and based on the credit value the multi-path cluster head routing can finally be found. The credit value is created and exchanged among the cluster heads only. Theoretical analysis combined with simulation results demonstrate that this protocol can save the resource, prolong the lifetime, and ensure the security and performance of the network.
基金Innovation Program of Shanghai Municipal Education Commission,China(No.12YZ191)
文摘Traditional clustering method is easy to slow convergence speed because of high data dimension and setting random initial clustering center. To improve these problems, a novel method combining subtractive clustering with fuzzy C-means( FCM)clustering will be advanced. In the method, the initial cluster number and cluster center can be obtained using subtractive clustering. On this basis,clustering result will be further optimized with FCM. In addition,the data dimension will be reduced through the analytic hierarchy process( AHP) before clustering calculating.In order to verify the effectiveness of fusion algorithm,an example about enterprise credit evaluation will be carried out. The results show that the fusion clustering algorithm is suitable for classifying high-dimension data,and the algorithm also does well in running up processing speed and improving visibility of result. So the method is suitable to promote the use.
文摘In this paper we applied the technique of Self Organizing Map (SOM) to segment individuals based on their credit information. SOM is an unsupervised machine learning method that reduces data complexity and dimensionality while keeping sits original topology, which is superior to other dimension reduction methods especially when features in data have unclear nonlinear relations. Through this method we provide more clear and intuitive segmentation that other traditional methods cannot achieve.
文摘众所周知,共识机制是区块链的核心,是区块链实现分布式存储的关键。随着各种区块链共识机制地出现,基于共识机制的优化方法也相继被提出,主要从优化共识过程以及控制共识节点的数量入手,解决共识机制吞吐量低、高时延、高资源等问题。然而,许多基于共识机制的优化缺乏理论的分析,也没有提及关键参数会影响共识机制的性能。为此,文中将以实用拜占庭算法(Practical Byzantine Fault Tolerance Algorithm,PBFT)、基于分组的实用拜占庭算法(Practical Byzantine Fault Tolerant Algorithm Based on Group,G-PBFT)以及基于分组和信誉的实用拜占庭算法(Practical Byzantine Fault Tolerant Algorithm Based on Clustering and Reputation,GR-PBFT)为例,构建三者的数学模型,进行性能分析。根据交易吞吐量、交易失败概率、区块认证失败概率和通信复杂度等性能指标进行对比。仿真结果表明:在同等节点数量下,G-PBFTD、GR-PBFT算法的吞吐量为PBFT的1.57倍、2.38倍;G-PBFTD、GR-PBFT算法的交易认证失败概率比PBFT下降了16%、39%;G-PBFTD、GR-PBFT算法的通信复杂度比PBFT下降了3.1倍、4.0倍,优化效果显著。
