技术领域Technical Field
本发明涉及区块链、隐私计算、绿色技术领域，更具体地，涉及一种基于区块链的绿色技术联合研发方法和系统。The present invention relates to the fields of blockchain, privacy computing, and green technology, and more specifically, to a green technology joint research and development method and system based on blockchain.
背景技术Background technique
区块链作为点对点网络、密码学、共享机制、智能合约等多种技术的集成创新，提供了一种在不可信网络中进行信息与价值传递交换的可信通道。区块链技术无论是在构建价值自由流通的互联网，还是在企业基于“建立联合多中心”的数据共享方面，都已成为全球炙手可热的概念，有着广阔的市场前景。As an integrated innovation of multiple technologies such as peer-to-peer networks, cryptography, sharing mechanisms, and smart contracts, blockchain provides a trusted channel for information and value transfer and exchange in untrusted networks. Blockchain technology has become a hot concept in the world, whether in building an Internet with free flow of value or in data sharing based on "establishing a joint multi-center" for enterprises, and has broad market prospects.
绿色技术是指遵循生态原理和生态经济规律，节约资源和能源，避免、消除或减轻生态环境污染和破坏，生态负效应最小的“无公害化”或“少公害化”的技术、工艺和产品的总称。其内容主要包括：污染控制和预防技术、源头削减技术、废物最少化技术、循环再生技术、生态工艺、绿色产品、净化技术等。可见绿色技术是一种与生态环境系统相协调的新型的现代技术系统。Green technology is a general term for technologies, processes and products that follow ecological principles and ecological economic laws, save resources and energy, avoid, eliminate or reduce ecological environmental pollution and damage, and minimize negative ecological effects, which are "pollution-free" or "pollution-less". Its contents mainly include: pollution control and prevention technology, source reduction technology, waste minimization technology, recycling technology, ecological process, green products, purification technology, etc. It can be seen that green technology is a new modern technology system that is coordinated with the ecological environment system.
区块链可作为绿色技术授权应用的基础，供合作各方之间构建基于隐私计算的联合研发系统，对合作各方来说，非本方的数据为“可用不可见”，这样就很大程度上保证了合作各方的数据安全。Blockchain can serve as the basis for the authorized application of green technology, allowing all parties to build a joint R&D system based on privacy computing. For all parties, data other than their own is "available but invisible", which largely ensures the data security of all parties.
但是，经发明人研究发现，在构建基于隐私计算的联合研发系统的过程中，区块链上的各方存在被伪装的可能，这就给基于区块链的绿色技术联合研发带来了很大的风险，需要构建可靠的识别算法来满足安全要求。However, the inventors have discovered that in the process of building a joint R&D system based on privacy computing, the parties on the blockchain may be disguised. This poses a great risk to the joint R&D of green technologies based on blockchain, and requires the construction of a reliable identification algorithm to meet security requirements.
发明内容Summary of the invention
针对区块链单节点伪装识别方法难以适应伪装联合研发团伙动态变化的问题，本发明提供一种基于区块链的绿色技术联合研发方法和系统，基于伪装联合研发团伙中节点社交模式之间的相似性和团伙社交模式之间的相似性，提出识别伪装联合研发团伙的新猜想和新范式，建立伪装联合研发团伙社交结构演化模式的表达方法，揭示团伙演化模式与团伙识别的关联关系，提升对伪装联合研发团伙中伪装节点精准识别的能力。In view of the problem that the blockchain single-node disguise identification method is difficult to adapt to the dynamic changes of disguised joint R&D groups, the present invention provides a green technology joint R&D method and system based on blockchain. Based on the similarity between the social modes of nodes in the disguised joint R&D group and the similarity between the social modes of groups, new conjectures and new paradigms for identifying disguised joint R&D groups are proposed, an expression method for the evolution mode of the social structure of disguised joint R&D groups is established, the correlation between the group evolution mode and group identification is revealed, and the ability to accurately identify disguised nodes in disguised joint R&D groups is improved.
为实现上述目的，本发明提供一种基于区块链的绿色技术联合研发方法和系统，包括以下步骤：To achieve the above objectives, the present invention provides a green technology joint research and development method and system based on blockchain, comprising the following steps:
步骤S1，构建区块链节点通讯记录集合；Step S1, constructing a blockchain node communication record set;
步骤S2，定义节点社交特征、潜在伪装联合研发团伙和团伙社交特征向量；Step S2, defining node social features, potential disguised joint R&D groups, and group social feature vectors;
步骤S3，基于预训练提取潜在伪装联合研发团伙；Step S3, extracting potential disguised joint R&D groups based on pre-training;
步骤S4，基于预训练的联合研发团伙内正常节点清洗；Step S4, cleaning normal nodes in the joint R&D group based on pre-training;
步骤S5，潜在伪装联合研发团伙社交行为特征提取；Step S5, extracting social behavior features of potential disguised joint R&D groups;
步骤S6，潜在伪装联合研发团伙社交结构子网络提取；Step S6, extracting the social structure sub-network of potential disguised joint R&D groups;
步骤S7，基于LSTM的伪装联合研发团伙识别；Step S7, identifying the disguised joint R&D group based on LSTM;
步骤S8，是否进行基于隐私计算的联合研发判定。Step S8: Determine whether to conduct joint research and development based on privacy computing.
进一步地，在所述步骤S1中，构建区块链节点通讯记录集合的具体步骤包括：给定集合U＝{u1,u2,...,uN}包含N个相互通讯的区块链节点，节点间的通讯记录集合表示为E＝{e1＝(u1,u2,t1,T1),e2＝(u2,u3,t2,T2),...,em＝(un,um,tm,Tm)}，其中e1＝(u1,u2,t1,T1)指的是访问节点u1在t1时刻和被访问节点u2进行通信，通信时长为T1。Furthermore, in the step S1, the specific steps of constructing a blockchain node communication record set include: a given set U = {u1 ,u2 ,...,uN } contains N blockchain nodes communicating with each other, and the communication record set between nodes is expressed as E = {e1 = (u1 ,u2 ,t1 ,T1 ), e2 = (u2 ,u3 ,t2 ,T2 ),..., em = (un ,um ,tm ,Tm )}, where e1 = (u1 ,u2 ,t1 ,T1 ) means that the visiting node u1 communicates with the visited node u2 at time t1 , and the communication duration is T1 .
进一步地，在所述步骤S2中，定义节点社交特征、潜在伪装联合研发团伙和团伙社交特征向量的具体步骤包括：将节点社交特征表示为特征矩阵Au＝{au1,au2,...,auN}，其中au代表每一个节点的特征向量，将集合G∈U表示为潜在伪装联合研发团伙集合，将向量aG表示为潜在伪装联合研发团伙G的团伙社交特征向量。Furthermore, in the step S2, the specific steps of defining node social characteristics, potential disguised joint R&D groups and group social feature vectors include: representing the node social characteristics as a feature matrix Au ={au1 ,au2 ,...,auN }, where au represents the feature vector of each node, representing the set G∈U as the set of potential disguised joint R&D groups, and representing the vector aG as the group social feature vector of the potential disguised joint R&D group G.
进一步地，在所述步骤S3中，基于预训练提取潜在伪装联合研发团伙的具体步骤包括：将在某段时间间隔Δt之内的特定节点ui接到通信的所有节点所构建的团伙Gi∈U作为潜在伪装联合研发团伙，即Gi中的节点u满足(u,ui,t,T)∈E，并且通讯时间差不超过Δt。Furthermore, in the step S3, the specific steps of extracting potential disguised joint R&D groups based on pre-training include: taking a groupGi∈U constructed by connecting a specific nodeui to all communicating nodes within a certain time interval Δt as a potential disguised joint R&D group, that is, the node u inGi satisfies (u, ui, t, T)∈E, and the communication time difference does not exceed Δt.
进一步地，在所述步骤S4中，基于预训练的联合研发团伙内正常节点清洗的具体步骤包括：Furthermore, in step S4, the specific steps of cleaning normal nodes in the joint R&D group based on pre-training include:
S41，利用提取的个体节点社交特征Au，并基于部分已知的正常节点标签学习有监督的逻辑回归模型进行预训练操作，将正常节点从潜在伪装联合研发团伙中分离，逻辑回归模型表示如下：S41, using the extracted individual node social featuresAu , and based on some known normal node labels, a supervised logistic regression model is learned for pre-training operation to separate normal nodes from potential disguised joint R&D groups. The logistic regression model is expressed as follows:
S42，在逻辑回归模型训练过程中，利用已知的部分伪装节点的真实标签数据，来计算该逻辑损失loss，表示如下：S42, during the training of the logistic regression model, the real label data of some known disguised nodes is used to calculate the logistic loss, which is expressed as follows:
其中m为已知的部分伪装节点数量，y为该节点是否为伪装节点标签，若是则y＝1，否则y＝0；Where m is the number of known partially disguised nodes, and y is the label of whether the node is a disguised node. If so, y = 1, otherwise y = 0;
S43，在基于个体社交行为模式使用逻辑回归模型划分出正常节点后，对得到的潜在伪装联合研发团伙的Gi进行清洗，在团伙中清洗掉社交行为明显正常的节点，最终得到潜在伪装联合研发团伙的候选团伙G。S43, after dividing the normal nodes using the logistic regression model based on the individual social behavior patterns, theGi of the potential disguised joint R&D group is cleaned, and the nodes with obviously normal social behaviors are cleaned out in the group, and finally the candidate group G of the potential disguised joint R&D group is obtained.
进一步地，在所述步骤S5中，潜在伪装联合研发团伙社交行为特征提取的具体步骤包括：Furthermore, in step S5, the specific steps of extracting social behavior features of potential disguised joint research and development groups include:
S51，设潜在伪装联合研发团伙的候选团伙G中包含节点集合(u1,u2,...,un)，将团伙G内每个节点社交行为特征au中的每一维特征值都分别计算其平均值和方差，得到该团伙社交行为特征向量aG，表示如下：S51, suppose that the candidate group G of the potential disguised joint R&D group contains a node set (u1 ,u2 ,...,un ), calculate the average value and variance of each dimension of the social behavior feature au of each node in the group G, and obtain the social behavior feature vector aG of the group, which is expressed as follows:
其中，表示节点u1的特征向量中的第一维值，avg()表示求平均操作，var()表示求方差操作；in, Represents the first dimension value in the feature vector of node u1 , avg() represents the average operation, and var() represents the variance operation;
S52，团伙社交行为特征向量的长度|aG|＝2×|au|为节点社交行为特征的二倍。用团伙社交行为特征向量的平均值表征团伙的社交共性，用团伙社交行为特征向量的方差表征团伙内每个节点的社交个性。S52, the length of the group social behavior feature vector |aG |=2×|au | is twice the length of the node social behavior feature. The average value of the group social behavior feature vector is used to represent the social commonality of the group, and the variance of the group social behavior feature vector is used to represent the social personality of each node in the group.
进一步地，在所述步骤S6中，潜在伪装联合研发团伙社交结构子网络提取的具体步骤包括：针对每一个基于规则方法提取的潜在伪装联合研发团伙，该团伙中的节点都共享一个相同的接收节点u1，针对该节点u1使用重启随机游走算法，提取一个s个节点所组成的子网络结构，表示如下：Furthermore, in step S6, the specific steps of extracting the social structure sub-network of the potential disguised joint R&D group include: for each potential disguised joint R&D group extracted based on the rule-based method, the nodes in the group share a same receiving node u1 , and a restarted random walk algorithm is used for the node u1 to extract a sub-network structure composed of s nodes, which is expressed as follows:
r＝cWr+(1-c)er＝cWr+(1-c)e
其中，r为提取的节点向量，令|r|＝s，便结束迭代循环，c为设置的参数，表示重启概率，e为起点节点，W为转移概率矩阵，由节点之间的通信记录表示，最终得到一个s个节点组成的小型社交网络图结构fnew＝{a1,a2,a3...as}，以表征潜在伪装联合研发团伙社交结构。Among them, r is the extracted node vector. Let |r| = s to end the iteration loop. c is the set parameter, which represents the restart probability. e is the starting node. W is the transition probability matrix, which is represented by the communication records between nodes. Finally, a small social network graph structure fnew = {a1 , a2 , a3 ... as } consisting of s nodes is obtained to characterize the social structure of the potential disguised joint R&D group.
进一步地，在所述步骤S7中，基于LSTM的伪装联合研发团伙识别的具体步骤包括：Furthermore, in step S7, the specific steps of identifying the disguised joint research and development group based on LSTM include:
S71，根据潜在伪装联合研发团伙社交结构fnew，对给定团队g，依次对特征项ai进行基于时间的间隔采样；S71, according to the potential disguised joint R&D group social structure fnew , for a given team g, perform time-based interval sampling on the feature items ai in turn;
S72，将时间划分为各个时间切片集合，表示为T＝{t1,t2,t3...tτ}，利用前述的LPC编码提取时间切片ti下基于特征ai的LPC谱特征通过拼接合并得到在时间切片ti下团队g的LPC谱特征：/>S72, divide the time into a set of time slices, represented as T = {t1 , t2 , t3 ... tτ }, and use the aforementioned LPC coding to extract the LPC spectrum features based on the features ai in the time slice ti By splicing and merging, we can get the LPC spectrum characteristics of team g at time sliceti :/>
S73，将团队g在各个时间切片下的LPC谱特征表示为并以此作为LSTM时序网络的输入，得到：S73, the LPC spectrum characteristics of team g at each time slice are expressed as And use this as the input of the LSTM timing network, and get:
式中ho和co分别表示模型输入的初始隐藏单元和状态单元，表示为LSTM网络的输出状态；Where ho and co represent the initial hidden unit and state unit of the model input respectively. Represented as the output state of the LSTM network;
S74，将作为团队g时序演化模式特征，实现团队g的身份识别预测：S74, will As the temporal evolution pattern feature of team g, the identity recognition prediction of team g is realized:
其中σ为sigmoid函数，如果yu＝1，将联合研发团伙识别为伪装联合研发团伙；如果yu＝0，将联合研发团伙识别为正常联合研发团伙。Where σ is a sigmoid function. Ifyu = 1, the joint research and development group is identified as a disguised joint research and development group; ifyu = 0, the joint research and development group is identified as a normal joint research and development group.
进一步地，在所述步骤S8中是否进行基于隐私计算的联合研发判定的具体步骤包括：如果联合研发团伙被识别为伪装联合研发团伙，则不进行基于隐私计算的联合研发，否则进行基于隐私计算的联合研发。Furthermore, in step S8, the specific steps of determining whether to conduct joint research and development based on privacy computing include: if the joint research and development group is identified as a disguised joint research and development group, joint research and development based on privacy computing is not conducted; otherwise, joint research and development based on privacy computing is conducted.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
图1为本发明实施例的流程图；FIG1 is a flow chart of an embodiment of the present invention;
图2为本发明步骤S4的分步流程图；FIG2 is a step-by-step flow chart of step S4 of the present invention;
图3为本发明步骤S5的分步流程图；FIG3 is a step-by-step flow chart of step S5 of the present invention;
图4为本发明步骤S7的分步流程图。FIG. 4 is a step-by-step flow chart of step S7 of the present invention.
具体实施方式Detailed ways
为了使本发明的目的和优点更加清楚明白，下面结合实施例对本发明作进一步描述；应当理解，此处所描述的具体实施例仅仅用于解释本发明，并不用于限定本发明。In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
下面参照附图来描述本发明的优选实施方式。本领域技术人员应当理解的是，这些实施方式仅仅用于解释本发明的技术原理，并非在限制本发明的保护范围。The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.
需要说明的是，在本发明的描述中，术语“上”、“下”、“左”、“右”、“内”、“外”等指示的方向或位置关系的术语是基于附图所示的方向或位置关系，这仅仅是为了便于描述，而不是指示或暗示所述装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.
此外，还需要说明的是，在本发明的描述中，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域技术人员而言，可根据具体情况理解上述术语在本发明中的具体含义。In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
请参阅图1所示，其为本发明实施例的流程图，本发明提供一种基于区块链的绿色技术联合研发方法和系统，包括：Please refer to FIG. 1 , which is a flow chart of an embodiment of the present invention. The present invention provides a green technology joint research and development method and system based on blockchain, including:
步骤S1，构建区块链节点通讯记录集合，具体而言，步骤S1包括：给定集合U＝{u1,u2,...,uN}包含N个相互通讯的区块链节点，节点间的通讯记录集合表示为E＝{e1＝(u1,u2,t1,T1),e2＝(u2,u3,t2,T2),...,em＝(un,um,tm,Tm)}，其中e1＝(u1,u2,t1,T1)指的是访问节点u1在t1时刻和被访问节点u2进行通信，通信时长为T1。Step S1, constructing a blockchain node communication record set. Specifically, step S1 includes: given a set U = {u1 ,u2 ,...,uN } containing N blockchain nodes communicating with each other, the communication record set between nodes is expressed as E = {e1 = (u1 ,u2 ,t1 ,T1 ), e2 = (u2 ,u3 ,t2 ,T2 ),...,em = (un ,um ,tm ,Tm )}, where e1 = (u1 ,u2 ,t1 ,T1 ) means that the visiting node u1 communicates with the visited node u2 at time t1 , and the communication duration is T1 .
步骤S2，定义节点社交特征、潜在伪装联合研发团伙和团伙社交特征向量。具体而言，步骤S2包括：将节点社交特征表示为特征矩阵Au＝{au1,au2,...,auN}，其中au代表每一个节点的特征向量，将集合G∈U表示为潜在伪装联合研发团伙集合，将向量aG表示为潜在伪装联合研发团伙G的团伙社交特征向量。Step S2, defining node social features, potential disguised joint R&D groups, and group social feature vectors. Specifically, step S2 includes: representing the node social features as a feature matrix Au ={au1 ,au2 ,...,auN }, where au represents the feature vector of each node, representing the set G∈U as the set of potential disguised joint R&D groups, and representing the vector aG as the group social feature vector of the potential disguised joint R&D group G.
步骤S3，基于预训练提取潜在伪装联合研发团伙，具体而言，步骤S3包括：将在某段时间间隔Δt之内的特定节点ui接到通信的所有节点所构建的团伙Gi∈U作为潜在伪装联合研发团伙，即Gi中的节点u满足(u,ui,t,T)∈E，并且通讯时间差不超过Δt。Step S3, extracting potential disguised joint R&D groups based on pre-training. Specifically, step S3 includes: taking the groupGi∈U constructed by connecting a specific nodeui to all communicating nodes within a certain time interval Δt as a potential disguised joint R&D group, that is, the node u inGi satisfies (u, ui, t, T)∈E, and the communication time difference does not exceed Δt.
步骤S4，基于预训练的联合研发团伙内正常节点清洗，具体而言，请参阅图2所示，步骤S4包括：Step S4, based on the pre-trained normal node cleaning in the joint R&D group, specifically, please refer to FIG. 2, step S4 includes:
S41，利用提取的个体节点社交特征Au，并基于部分已知的正常节点标签学习有监督的逻辑回归模型进行预训练操作，将正常节点从潜在伪装联合研发团伙中分离，逻辑回归模型表示如下：S41, using the extracted individual node social featuresAu , and based on some known normal node labels, a supervised logistic regression model is learned for pre-training operation to separate normal nodes from potential disguised joint R&D groups. The logistic regression model is expressed as follows:
S42，在逻辑回归模型训练过程中，利用已知的部分伪装节点的真实标签数据，来计算该逻辑损失loss，表示如下：S42, during the training of the logistic regression model, the real label data of some known disguised nodes is used to calculate the logistic loss, which is expressed as follows:
其中m为已知的部分伪装节点数量，y为该节点是否为伪装节点标签，若是则y＝1，否则y＝0；Where m is the number of known partially disguised nodes, and y is the label of whether the node is a disguised node. If so, y = 1, otherwise y = 0;
S43，在基于个体社交行为模式使用逻辑回归模型划分出正常节点后，对得到的潜在伪装联合研发团伙的Gi进行清洗，在团伙中清洗掉社交行为明显正常的节点，最终得到潜在伪装联合研发团伙的候选团伙G。S43, after dividing the normal nodes using the logistic regression model based on the individual social behavior patterns, theGi of the potential disguised joint R&D group is cleaned, and the nodes with obviously normal social behaviors are cleaned out in the group, and finally the candidate group G of the potential disguised joint R&D group is obtained.
步骤S5，潜在伪装联合研发团伙社交行为特征提取。基于潜在伪装联合研发团伙节点社交行为特征，计算并合并出伪装联合研发团伙社交行为特征，以用来识别真实伪装联合研发团伙，用于同时表征团伙社交行为模式中的团伙的共性特征和节点的个性特征，具体而言，请参阅图3所示，步骤S5包括：Step S5, extracting social behavior features of potential disguised joint R&D groups. Based on the social behavior features of potential disguised joint R&D group nodes, calculate and merge the social behavior features of disguised joint R&D groups to identify real disguised joint R&D groups, and to simultaneously characterize the common features of the groups and the individual features of the nodes in the group social behavior pattern. Specifically, please refer to FIG3 , step S5 includes:
S51，设潜在伪装联合研发团伙的候选团伙G中包含节点集合(u1,u2,...,un)，将团伙G内每个节点社交行为特征au中的每一维特征值都分别计算其平均值和方差，得到该团伙社交行为特征向量aG，表示如下：S51, suppose that the candidate group G of the potential disguised joint R&D group contains a node set (u1 ,u2 ,...,un ), calculate the average value and variance of each dimension of the social behavior feature au of each node in the group G, and obtain the social behavior feature vector aG of the group, which is expressed as follows:
其中，表示节点u1的特征向量中的第一维值，avg()表示求平均操作，var()表示求方差操作；in, Represents the first dimension value in the feature vector of node u1 , avg() represents the average operation, and var() represents the variance operation;
S52，团伙社交行为特征向量的长度|aG|＝2×|au|为节点社交行为特征的二倍，用团伙社交行为特征向量的平均值表征团伙的社交共性，用团伙社交行为特征向量的方差表征团伙内每个节点的社交个性。S52, the length of the group social behavior feature vector |aG |=2×|au | is twice the social behavior feature of the node. The average value of the group social behavior feature vector is used to represent the social commonality of the group, and the variance of the group social behavior feature vector is used to represent the social personality of each node in the group.
步骤S6，潜在伪装联合研发团伙社交结构子网络提取。由于潜在伪装联合研发团伙的社交行为模式与正常联合研发团队的差异性也同样体现在社交网络的图结构中，因此提取每一个潜在伪装联合研发团伙社交结构子网络来辅助挖掘伪装团伙。具体而言，步骤S6包括：针对每一个基于规则方法提取的潜在伪装联合研发团伙，该团伙中的节点都共享一个相同的接收节点u1，针对该节点u1使用重启随机游走算法，提取一个s个节点所组成的子网络结构，表示如下：Step S6, extracting the social structure sub-network of potential disguised joint R&D groups. Since the difference between the social behavior patterns of potential disguised joint R&D groups and normal joint R&D teams is also reflected in the graph structure of the social network, each potential disguised joint R&D group social structure sub-network is extracted to assist in mining disguised groups. Specifically, step S6 includes: for each potential disguised joint R&D group extracted based on the rule-based method, the nodes in the group share a same receiving node u1 , and a restart random walk algorithm is used for the node u1 to extract a sub-network structure composed of s nodes, which is expressed as follows:
r＝cWr+(1-c)er＝cWr+(1-c)e
其中，r为提取的节点向量，令|r|＝s，便结束迭代循环，c为设置的参数，表示重启概率，e为起点节点，W为转移概率矩阵，由节点之间的通信记录表示。Among them, r is the extracted node vector, let |r| = s, then the iteration loop ends, c is the set parameter, indicating the restart probability, e is the starting node, and W is the transition probability matrix, which is represented by the communication records between nodes.
最终得到一个s个节点组成的小型社交网络图结构fnew＝{a1,a2,a3...as}，以表征潜在伪装联合研发团伙社交结构。Finally, a small social network graph structure fnew ={a1 ,a2 ,a3 ...as } consisting of s nodes is obtained to characterize the social structure of the potential disguised joint R&D group.
步骤S7，基于LSTM的伪装联合研发团伙识别，具体而言，请参阅图4所示，步骤S7包括：Step S7, LSTM-based identification of disguised joint research and development groups, specifically, please refer to FIG. 4, step S7 includes:
S71，根据潜在伪装联合研发团伙社交结构fnew，对给定团队g，依次对特征项ai进行基于时间的间隔采样；S71, according to the potential disguised joint R&D group social structure fnew , for a given team g, perform time-based interval sampling on the feature items ai in turn;
S72，将时间划分为各个时间切片集合，表示为T＝{t1,t2,t3...tτ}，利用前述的LPC编码提取时间切片ti下基于特征ai的LPC谱特征通过拼接合并得到在时间切片ti下团队g的LPC谱特征：/>S72, divide the time into a set of time slices, represented as T = {t1 , t2 , t3 ... tτ }, and use the aforementioned LPC coding to extract the LPC spectrum features based on the features ai in the time slice ti By splicing and merging, we can get the LPC spectrum characteristics of team g at time sliceti :/>
S73，将团队g在各个时间切片下的LPC谱特征表示为并以此作为LSTM时序网络的输入，得到：S73, the LPC spectrum characteristics of team g at each time slice are expressed as And use this as the input of the LSTM timing network, and get:
式中ho和co分别表示模型输入的初始隐藏单元和状态单元，表示为LSTM网络的输出状态；Where ho and co represent the initial hidden unit and state unit of the model input respectively. Represented as the output state of the LSTM network;
S74，将作为团队g时序演化模式特征，实现团队g的身份识别预测：S74, will As the temporal evolution pattern feature of team g, the identity recognition prediction of team g is realized:
其中σ为sigmoid函数，如果yu＝1，将联合研发团伙识别为伪装联合研发团伙；如果yu＝0，将联合研发团伙识别为正常联合研发团伙。Where σ is a sigmoid function. Ifyu = 1, the joint research and development group is identified as a disguised joint research and development group; ifyu = 0, the joint research and development group is identified as a normal joint research and development group.
步骤S8，是否进行基于隐私计算的联合研发判定，具体而言，步骤S8包括：如果联合研发团伙被识别为伪装联合研发团伙，则不进行基于隐私计算的联合研发，否则进行基于隐私计算的联合研发。Step S8, whether to conduct joint research and development based on privacy computing. Specifically, step S8 includes: if the joint research and development group is identified as a disguised joint research and development group, joint research and development based on privacy computing will not be conducted; otherwise, joint research and development based on privacy computing will be conducted.
至此，已经结合附图所示的优选实施方式描述了本发明的技术方案，但是，本领域技术人员容易理解的是，本发明的保护范围显然不局限于这些具体实施方式。在不偏离本发明的原理的前提下，本领域技术人员可以对相关技术特征做出等同的更改或替换，这些更改或替换之后的技术方案都将落入本发明的保护范围之内。So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
以上所述仅为本发明的优选实施例，并不用于限制本发明；对于本领域的技术人员来说，本发明可以有各种更改和变化。凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410140666.0ACN117978497B (en) | 2024-02-01 | 2024-02-01 | A green technology joint research and development method and system based on blockchain |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410140666.0ACN117978497B (en) | 2024-02-01 | 2024-02-01 | A green technology joint research and development method and system based on blockchain |
| Publication Number | Publication Date |
|---|---|
| CN117978497Atrue CN117978497A (en) | 2024-05-03 |
| CN117978497B CN117978497B (en) | 2024-08-23 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410140666.0AActiveCN117978497B (en) | 2024-02-01 | 2024-02-01 | A green technology joint research and development method and system based on blockchain |
| Country | Link |
|---|---|
| CN (1) | CN117978497B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111861756A (en)* | 2020-08-05 | 2020-10-30 | 哈尔滨工业大学（威海） | A kind of gang detection method based on financial transaction network and its realization device |
| CN113129028A (en)* | 2020-01-10 | 2021-07-16 | 联洋国融(北京)科技有限公司 | Rogue user detection system based on time sequence neural network model |
| US20230177512A1 (en)* | 2021-12-08 | 2023-06-08 | Chime Financial, Inc. | Generating a fraud prediction utilizing a fraud-prediction machine-learning model |
| CN116822569A (en)* | 2022-03-18 | 2023-09-29 | 中移(上海)信息通信科技有限公司 | Model training method, device, related equipment and storage medium |
| CN117035082A (en)* | 2023-10-10 | 2023-11-10 | 北京江融信科技有限公司 | Knowledge graph real-time operation method and system for identifying fraudulent party |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113129028A (en)* | 2020-01-10 | 2021-07-16 | 联洋国融(北京)科技有限公司 | Rogue user detection system based on time sequence neural network model |
| CN111861756A (en)* | 2020-08-05 | 2020-10-30 | 哈尔滨工业大学（威海） | A kind of gang detection method based on financial transaction network and its realization device |
| US20230177512A1 (en)* | 2021-12-08 | 2023-06-08 | Chime Financial, Inc. | Generating a fraud prediction utilizing a fraud-prediction machine-learning model |
| CN116822569A (en)* | 2022-03-18 | 2023-09-29 | 中移(上海)信息通信科技有限公司 | Model training method, device, related equipment and storage medium |
| CN117035082A (en)* | 2023-10-10 | 2023-11-10 | 北京江融信科技有限公司 | Knowledge graph real-time operation method and system for identifying fraudulent party |
| Title |
|---|
| 邵蓥侠;冯是聪;: "社交网络分析在公共安全领域的应用", 大数据, no. 02, 20 March 2017 (2017-03-20)* |
| Publication number | Publication date |
|---|---|
| CN117978497B (en) | 2024-08-23 |
| Publication | Publication Date | Title |
|---|---|---|
| Cheng et al. | Leveraging semisupervised hierarchical stacking temporal convolutional network for anomaly detection in IoT communication | |
| Liu et al. | Keep your data locally: Federated-learning-based data privacy preservation in edge computing | |
| Chen et al. | Privacy preserving point-of-interest recommendation using decentralized matrix factorization | |
| CN112464292B (en) | Method and device for training graph neural network based on privacy protection | |
| Li et al. | Online federated multitask learning | |
| CN112395466B (en) | Fraud node identification method based on graph embedded representation and cyclic neural network | |
| WO2021164365A1 (en) | Graph neural network model training method, apparatus and system | |
| Hossain et al. | Active deep learning for activity recognition with context aware annotator selection | |
| CN117729027A (en) | Abnormal behavior detection method, device, electronic equipment and storage medium | |
| Ji et al. | Multi-range gated graph neural network for telecommunication fraud detection | |
| WO2020038100A1 (en) | Feature relationship recommendation method and apparatus, computing device and storage medium | |
| CN114048328B (en) | Knowledge-graph link prediction method and system based on conversion hypothesis and message transmission | |
| Lei et al. | Federated learning over coupled graphs | |
| CN114036405A (en) | A method and system for social recommendation based on graph convolutional network | |
| CN112446556B (en) | Communication network user calling object prediction method based on expression learning and behavior characteristics | |
| Qi et al. | Privacy-preserving cross-area traffic forecasting in ITS: A transferable spatial-temporal graph neural network approach | |
| CN115292585A (en) | A Conversational Item Recommendation Method Based on Context-Aware Graph Neural Networks | |
| CN114254738A (en) | Construction method and application of dynamic graph convolutional neural network model with two-layer evolution | |
| CN116628911A (en) | Trust Evaluation Method Based on Graph Neural Network | |
| Zhou et al. | Learning to correlate accounts across online social networks: An embedding-based approach | |
| Sah et al. | Aggregation techniques in federated learning: Comprehensive survey, challenges and opportunities | |
| Cao et al. | Implicit user relationships across sessions enhanced graph for session-based recommendation | |
| CN117978497B (en) | A green technology joint research and development method and system based on blockchain | |
| Yao | A propagation model for integrating web of things and social networks | |
| CN116467415A (en) | Bidirectional cross-domain conversational recommendation method based on GCNsformer hybrid network and multi-channel semantics |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |