Industrial Internet Ecosystem Evolution Prediction Method and Device Based on Link Prediction

Abstract

The present application relates to a method and device for predicting the evolution of an industrial Internet ecosystem based on link prediction. The method includes: determining an industrial collaboration network sequence model and a formal definition of the problem; based on the industrial collaboration network sequence model and the formal definition of the problem, extracting at least one industrial collaboration network evolution feature and verifying at least one industrial collaboration network evolution feature to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; predicting the evolution process of the industrial Internet ecosystem according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction. Thus, problems such as the high threshold for leading enterprises to build an industrial ecological community and the difficulty in supporting the long-term and sustainable development of the entire industrial Internet ecosystem are solved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a method and device for predicting the evolution of the industrial Internet ecosystem based on link prediction. Background Art

[0002] An ecosystem is considered a complex system that is robust, scalable, continuous, dynamic, adaptive, self-organizing, and capable of automatically solving complex dynamic problems. When the concept of an ecosystem is borrowed into the industrial Internet, the concept of an industrial Internet ecosystem is formed. In the industrial Internet ecosystem, a large number of new enterprises continuously enter the system, enriching the functions that the entire system can provide to consumers. Among them, some enterprises have industry competitive advantages or leadership capabilities, and under certain conditions, they can cooperate with upstream and downstream enterprises to form an industrial ecological community that meets the specific needs of consumers. However, such a community does not always occur. There are many reasons for this phenomenon, but from the perspective of the ecosystem itself, the main reasons include the following aspects:

[0003] (1) Although many enterprises have joined the ecosystem, they have not introduced actual business, and at the same time, consumer demands are constantly changing and difficult to accurately grasp. Under such circumstances with many uncertain factors, leading enterprises have no incentive to form relatively stable industrial ecological communities;

[0004] (2) A large number of enterprises bring the problem of information overload for selection. There are a large number of enterprises in each link of the supply chain that can provide corresponding functions, thus bringing difficulties for selection. A corresponding recommendation method is needed to recommend the most suitable objects for leading enterprises from a large number of enterprises, reducing the difficulties for leading enterprises to form industrial ecological communities;

[0005] (3) A large number of enterprises compete with each other in the ecosystem. Among them, the enterprises that fail in the competition may withdraw from the ecosystem, and the withdrawal of these enterprises will affect the overall operating efficiency of the industrial ecological community to which they belong. To ensure the long-term development of the industrial ecological community, it is necessary to filter out the enterprises that may withdraw at the initial stage of community establishment, improving the lifespan of the constructed ecological community.

[0006] Therefore, currently, the threshold for leading enterprises to form industrial ecological communities is relatively high, making it difficult to support the long-term and continuous development of the entire industrial Internet ecosystem, which urgently needs to be solved. Summary of the Invention

[0007] This application provides a method and device for predicting the evolution of the industrial Internet ecosystem based on link prediction to solve problems such as the relatively high threshold for leading enterprises to form industrial ecological communities and the difficulty in supporting the long-term and continuous development of the entire industrial Internet ecosystem.

[0008] The first aspect of the embodiments of the present application provides an industrial Internet ecological evolution prediction method based on link prediction, including the following steps: determining an industrial collaboration network sequence model and a formal definition of the problem; based on the industrial collaboration network sequence model and the formal definition of the problem, extracting at least one industrial collaboration network evolution feature and verifying the at least one industrial collaboration network evolution feature to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; and predicting the industrial Internet ecological evolution process according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction.

[0009] Optionally, in an embodiment of the present application, the determining the industrial collaboration network sequence model and the formal definition of the problem includes: constructing a collaboration relationship based on the organization mode of the ecological community formed within the target time period to generate a fragmented industrial collaboration network; obtaining a fragmented industrial collaboration network sequence from the fragmented industrial collaboration networks of each time slice of the target time series; generating a past industrial collaboration network and a forward industrial collaboration network based on the fragmented industrial collaboration network sequence; obtaining a past industrial collaboration matrix and a forward industrial collaboration matrix according to the past industrial collaboration network and the forward industrial collaboration network; predicting the network of the fragmented industrial collaboration network sequence in the future, and based on the past industrial collaboration network and the forward industrial collaboration network, while forming the industrial collaboration network sequence model, determining the formal definition of the problem.

[0010] Optionally, in an embodiment of the present application, the past industrial collaboration matrix is expressed as:

[0011]

[0012] where t i is the i-th time slice, l is the time window, θ is the time decay parameter, represents the fragmented industrial collaboration matrix;

[0013] And, the forward industrial collaboration matrix is expressed as:

[0014]

[0015] where g is the forward time window.

[0016] Optionally, in an embodiment of the present application, the determining the formal definition of the problem includes: based on the given fragmented industrial collaboration network sequence, past time window, time decay parameter, and forward time window, constructing the corresponding past industrial collaboration network and the corresponding forward industrial collaboration network, and finding a mapping function to calculate the probability of occurrence in the corresponding forward industrial collaboration network for any pair of nodes in the corresponding past industrial collaboration network.

[0017] Optionally, in an embodiment of the present application, predicting the industrial Internet ecological evolution process based on the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction includes: analyzing, based on a preset unsupervised linear classifier, the degree to which at least one indicator reflects the evolution characteristics of the industrial collaboration network on the evolution process of the industrial collaboration network; quantifying the degree of reflection, and using the ranking of the feature values corresponding to the nodes as an indicator for node evolution prediction.

[0018] An embodiment of the second aspect of the present application provides an industrial Internet ecological evolution prediction device based on link prediction, including: a definition module for determining an industrial collaboration network sequence model and a formal definition of the problem; a verification module for extracting at least one industrial collaboration network evolution feature based on the industrial collaboration network sequence model and the formal definition of the problem, and verifying the at least one industrial collaboration network evolution feature to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; and a prediction module for predicting the industrial Internet ecological evolution process based on the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction.

[0019] Optionally, in an embodiment of the present application, the definition module includes: a first generation unit for constructing a collaboration relationship based on the organization mode of the ecological community formed within a target time period to generate a fragmented industrial collaboration network; a processing unit for obtaining a fragmented industrial collaboration network sequence from the fragmented industrial collaboration networks of each time slice of the target time series; a second generation unit for generating a past industrial collaboration network and a forward industrial collaboration network based on the fragmented industrial collaboration network sequence; an acquisition unit for obtaining a past industrial collaboration matrix and a forward industrial collaboration matrix according to the past industrial collaboration network and the forward industrial collaboration network; and a modeling unit for predicting the network of the fragmented industrial collaboration network sequence in the future, and forming the industrial collaboration network sequence model based on the past industrial collaboration network and the forward industrial collaboration network, while determining the formal definition of the problem.

[0020] Optionally, in an embodiment of the present application, the past industrial collaboration matrix is expressed as:

[0021]

[0022] where t i is the i-th time slice, l is the time window, θ is the time decay parameter, represents the fragmented industrial collaboration matrix;

[0023] And the forward industrial collaboration matrix is expressed as:

[0024]

[0025] Among them, g is the forward time window.

[0026] Optionally, in an embodiment of the present application, the modeling unit is specifically configured to construct a corresponding past industrial collaboration network and a corresponding forward industrial collaboration network based on a given sharded industrial collaboration network sequence, a past time window, a time decay parameter, and a forward time window, and find a mapping function to calculate the probability of occurrence in the corresponding forward industrial collaboration network for any pair of nodes in the corresponding past industrial collaboration network.

[0027] Optionally, in an embodiment of the present application, the prediction module includes: an analysis unit for analyzing the degree to which at least one metric reflects the evolution characteristics of the industrial collaboration network in the evolution process of the industrial collaboration network based on a preset unsupervised linear classifier; a quantization unit for quantifying the degree of reflection, and using the ranking of the eigenvalue corresponding to the node as an indicator for predicting the evolution of the node pair.

[0028] An embodiment of the third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the industrial Internet ecological evolution prediction method based on link prediction as described in the above embodiments.

[0029] An embodiment of the fourth aspect of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the program is executed by a processor, it implements the industrial Internet ecological evolution prediction method based on link prediction as above.

[0030] Therefore, the embodiments of the present application have the following beneficial effects:

[0031] The embodiments of the present application can determine the industrial collaboration network sequence model and the formal definition of the problem; based on the industrial collaboration network sequence model and the formal definition of the problem, extract at least one industrial collaboration network evolution characteristic, and verify at least one industrial collaboration network evolution characteristic to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; predict the industrial Internet ecological evolution process according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction, so as to effectively predict the evolution of the industrial collaboration network according to the historical industrial collaboration network, and extract successful ecological community organization methods from it to support the recommendation of cooperative enterprises. Thus, it solves the problems that the threshold for leading enterprises to build industrial ecological communities is relatively high and it is difficult to support the long-term and sustainable development of the entire industrial Internet ecosystem.

[0032] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present application. Brief Description of the Drawings

[0033] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the drawings, where:

[0034] Figure 1 It is a flowchart of a method for predicting the evolution of an industrial Internet ecosystem based on link prediction provided according to an embodiment of the present application;

[0035] Figure 2 It is a schematic diagram of the relationship between three industrial collaboration networks provided according to an embodiment of the present application;

[0036] Figure 3 It is a schematic diagram of the evolution characteristics of a two-dimensional industrial collaboration network provided according to an embodiment of the present application;

[0037] Figure 4 It is a schematic diagram of the effect analysis of eigenvalue on the prediction of the evolution of a collaboration network (taking i = 86, l = 10, g = 1, θ = 0 as an example) provided according to an embodiment of the present application;

[0038] Figure 5 It is an example diagram of an apparatus for predicting the evolution of an industrial Internet ecosystem based on link prediction according to an embodiment of the present application;

[0039] Figure 6 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application.

[0040] Among them, 10 - an apparatus for predicting the evolution of an industrial Internet ecosystem based on link prediction, 100 - a definition module, 200 - a verification module, 300 - a prediction module, 601 - a memory, 602 - a processor, 603 - a communication interface. Detailed Description of the Embodiments

[0041] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present application, and should not be construed as limiting the present application.

[0042] The method and device for predicting the evolution of the industrial Internet ecosystem based on link prediction according to the embodiments of the present application will be described below with reference to the accompanying drawings. In view of the problems mentioned in the above background art, the present application provides a method for predicting the evolution of the industrial Internet ecosystem based on link prediction. In this method, a sequence model of the industrial collaboration network and a formal definition of the problem are determined; based on the sequence model of the industrial collaboration network and the formal definition of the problem, at least one evolution feature of the industrial collaboration network is extracted and at least one evolution feature of the industrial collaboration network is verified to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; the evolution process of the industrial Internet ecosystem is predicted according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction, so as to effectively predict the evolution of the industrial collaboration network according to the historical industrial collaboration network, and extract the successful ecological community organization methods from it to support the recommendation of partner enterprises. Thus, the problems that the threshold for leading enterprises to build industrial ecological communities is relatively high and it is difficult to support the long-term and sustainable development of the entire industrial Internet ecosystem are solved.

[0043] Specifically, Figure 1 is a flowchart of a method for predicting the evolution of the industrial Internet ecosystem based on link prediction provided by an embodiment of the present application.

[0044] As Figure 1 shown, the method for predicting the evolution of the industrial Internet ecosystem based on link prediction includes the following steps:

[0045] In step S101, a sequence model of the industrial collaboration network and a formal definition of the problem are determined.

[0046] Those skilled in the art should understand that the prediction of the evolution of the industrial collaboration network can effectively reflect the evolution process of the ecological community organization method in the industrial Internet ecosystem. Therefore, in the embodiments of the present application, the relevant models and formal definitions of the problems of the industrial collaboration network can be defined first, so as to predict and analyze the evolution process of the industrial collaboration network.

[0047] Optionally, in an embodiment of the present application, determining the sequence model of the industrial collaboration network and the formal definition of the problem includes: constructing a collaboration relationship based on the organization method of the ecological community formed within the target time period to generate a fragmented industrial collaboration network; obtaining a fragmented industrial collaboration network sequence from the fragmented industrial collaboration networks of each time slice of the target time series; generating a past industrial collaboration network and a forward industrial collaboration network based on the fragmented industrial collaboration network sequence; obtaining a past industrial collaboration matrix and a forward industrial collaboration matrix according to the past industrial collaboration network and the forward industrial collaboration network; predicting the network of the fragmented industrial collaboration network sequence in the future, and based on the past industrial collaboration network and the forward industrial collaboration network, while forming a sequence model of the industrial collaboration network, determining the formal definition of the problem.

[0048] It should be noted that the specific steps for defining the relevant models and problem formalization of the industrial collaboration network in the embodiments of this application are as follows:

[0049] 1. Define the segmented industrial collaboration network:

[0050] Given the i-th time segment t i , obtain the organizational form of the ecological community formed during this time period, construct its collaboration relationship, and further generate an industrial collaboration network. Since this industrial collaboration network only reflects the collaboration relationship within a certain time segment, it is defined as the segmented industrial collaboration network

[0051] 2. Define the sequence of segmented industrial collaboration networks:

[0052] From the definition of the segmented industrial collaboration network, it can be known that each segmented industrial collaboration network describes the collaboration mode during this time period. Therefore, the sequence of segmented industrial collaboration networks can be easily defined as follows:

[0053] Given a time sequence t1, K t k , for each time segment t i , 1 ≤ i ≤ k, obtain its segmented industrial collaboration network Obtain the sequence of segmented industrial collaboration networks where k represents the length of the time sequence, thereby describing the evolution process of the collaboration mode in the industrial Internet ecosystem by defining the sequence of segmented industrial collaboration networks;

[0054] 3. Define the past industrial collaboration network and the forward industrial collaboration network:

[0055] Given the i-th time segment t i , and the time window l, obtain the organizational form of the ecological community during the period from the max(1, i - l + 1)-th time segment to the i-th time segment, and generate the corresponding industrial collaboration network. Since this industrial collaboration network covers the industrial collaboration relationships in l time segments up to t i , it is defined as the past industrial collaboration network

[0056] Given the i-th time segment t i , and the forward time window g, obtain the organizational form of the ecological community during the period from the i-th time segment to the i + g-th time segment, and generate the corresponding industrial collaboration network. Since this industrial collaboration network reflects the subsequent changes in the industrial collaboration network, it is defined as the forward industrial collaboration network

[0057] 4. Define the past industrial collaboration matrix and the forward industrial collaboration matrix:

[0058] The sharded industrial collaboration network can be described by the sharded industrial collaboration matrix The industrial collaboration matrix of the past collaboration network can be defined as The forward industrial collaboration network is Both the past collaboration network and the forward industrial collaboration network can be obtained by aggregating the sharded industrial collaboration network;

[0059] 5. Formal definition of the problem of predicting the sequential evolution of the industrial collaboration network:

[0060] Since the evolution process of the industrial collaboration network can be described by a sequence of sharded industrial collaboration networks, the problem of predicting the industrial collaboration pattern can be transformed into the problem of predicting the sequence of industrial collaboration networks.

[0061] Thus, the embodiments of this application can, by given a sequence of sharded industrial collaboration networks and find a mapping according to the following formula to predict the network in the future of the sequence of sharded industrial collaboration networks;

[0062]

[0063] Based on the definitions of the past industrial collaboration network and the forward industrial collaboration network, the embodiments of this application can further define it as a mapping problem between the past industrial collaboration network and the forward industrial collaboration network, as described below:

[0064] Given a sequence of sharded industrial collaboration networks The past time window l, the time decay parameter θ, the forward time window g, construct the past industrial collaboration network and the forward industrial collaboration network Find the mapping function Realize the mapping of the past industrial collaboration network to the forward industrial collaboration network.

[0065] According to the status of enterprise nodes and edges in the nodes in can be divided into two types: reused nodes and cold-start nodes; Since the cold-start nodes do not contain corresponding information in and the proportion is small, the embodiments of this application do not analyze them further;

[0066] Thus, the above problem is transformed into the problem of predicting the reused enterprise node edges (RSC) and the emerging enterprise node edges (ESC), and further realizes the formal definition of the problem of predicting the sequential evolution of the industrial collaboration network.

[0067] Optionally, in an embodiment of the present application, the past industrial collaboration matrix is represented as:

[0068]

[0069] where t i is the i-th time slice, l is the time window, θ is the time decay parameter, represents the sliced industrial collaboration matrix;

[0070] And, the forward industrial collaboration matrix is represented as:

[0071]

[0072] where g is the forward time window.

[0073] It should be noted that, in order to describe the differences in the sliced industrial collaboration networks in different time periods, the time decay parameter θ is further defined. Therefore, in the embodiments of the present application, the industrial collaboration matrix of the past collaboration network can be defined as:

[0074]

[0075] where, if θ = 0, it means that the impacts of all sliced industrial collaboration networks on the past industrial collaboration network are the same; if 0 < θ < 1, it means that the organizational methods of ecological communities in more distant past have less impact on the past industrial collaboration network. Therefore, the larger θ is, the more obvious this decay effect is; if the time window is defined as l = ∞, at this time all historical slices will be covered, and the past industrial network then represents the industrial collaboration network composed of all historical ecological community organizations

[0076] Similarly, the industrial collaboration matrix of the forward collaboration network can be defined as:

[0077]

[0078] Figure 2 Shows the relationship between the sliced industrial collaboration network, the past industrial collaboration network, and the forward industrial collaboration network in the embodiments of the present application. Thus, the embodiments of the present application provide a basis for the formal definition of the industrial collaboration network sequence evolution prediction problem by defining the past industrial collaboration matrix and the forward industrial collaboration matrix.

[0079] Optionally, in an embodiment of the present application, determining a formal definition of the problem includes: based on a given sharded industrial collaboration network sequence, a past time window, a time decay parameter, and a forward time window, constructing a corresponding past industrial collaboration network and a corresponding forward industrial collaboration network, and finding a mapping function to calculate the probability of occurrence in the corresponding forward industrial collaboration network for any pair of nodes in the corresponding past industrial collaboration network.

[0080] Those skilled in the art should understand that in the industrial Internet ecosystem, the industrial collaboration network reflects the collaboration relationships between enterprises. During the long-term evolution process, the mutual collaboration among a large number of enterprises has formed complex correlation relationships, showing the characteristics of complex network systems, such as small-world, power-law characteristics, and disassortativity, etc. At the same time, the industrial collaboration network has predictability, that is, the changes in the future collaboration network can be predicted through the analysis of historical collaboration relationships. Since the industrial collaboration network actually represents the collaboration history between enterprises and reflects the historical needs of consumers, the evolution process of the industrial collaboration network actually also reflects the change process of the needs of enterprises and consumers in the system. Based on this, the problem can be stated as: given the historical industrial collaboration network, how to effectively predict the evolution of the industrial collaboration network and extract the successful ecological community organization methods from it to support the recommendation of cooperative enterprises.

[0081] Therefore, the embodiment of the present application can formally define the problem from the perspective of link prediction as follows:

[0082] Given a sharded industrial collaboration network sequence A past time window l, a time decay parameter θ, a forward time window g, constructing a past industrial collaboration network And a forward industrial collaboration network Finding a mapping function For Any pair of nodes (s u , s v ) in it, calculate the probability of its occurrence in It.

[0083] Thus, the embodiment of the present application provides a basis and guidance for the subsequent predictive analysis of the evolution of the industrial Internet ecosystem by defining the network sequence model and the formalization of the problem.

[0084] In step S102, based on the industrial collaboration network sequence model and the formal definition of the problem, at least one industrial collaboration network evolution feature is extracted, and at least one industrial collaboration network evolution feature is verified to obtain the predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction.

[0085] After determining the industrial collaboration network sequence model and the formal definition of the problem, further, the embodiments of the present application can extract and verify the evolutionary characteristics of the industrial collaboration network, so as to predict and analyze the evolution of the industrial collaboration network from the perspective of network prediction.

[0086] Specifically, the specific process of extracting the evolutionary characteristics of the industrial collaboration network in the embodiments of the present application is as follows:

[0087] 1. Definition of node pair features

[0088] The characteristics of a node pair refer to defining relevant topological and temporal characteristics with the two nodes in the node pair as a whole, mainly including the following four aspects of characteristics:

[0089] 1) Weight of the edge connecting the node pair (Current Sum Edge Weight, CSEW)

[0090] The cumulative distribution of the edge weights in the industrial collaboration network satisfies the power-law distribution, and this power-law effect is gradually strengthening. Therefore, the higher the weight of the edge connecting two enterprise nodes, the higher the probability that these two enterprise nodes will collaborate in the same ecological community in the future and thus appear in the forward industrial collaboration network. Therefore, the weight w i (s u , s v ) of the edge connecting the node pair will be used as a feature;

[0091] Among them, (s u , s v ) represents the edge formed by two nodes in the collaboration network, and s u , s v represents the nodes in the collaboration network;

[0092] 2) Weight of common adjacent nodes (Weighted Common Neighbors, WCN)

[0093] The industrial collaboration network has obvious small-world characteristics, that is, a short characteristic path length and a high clustering coefficient. Therefore, there is a higher possibility that the adjacent enterprise nodes of each enterprise node will establish connections in the future; from the perspective of enterprise nodes, that is, the more common adjacent nodes these two enterprise nodes have, the higher the probability that these two enterprise nodes will build collaborative edges in the future; since the industrial collaboration network is a weighted network, the following feature of the weight of common adjacent nodes is defined:

[0094]

[0095] Among them, s k ∈Ne i (s u )I Nei (s v ) represents the common adjacent nodes between two enterprise nodes;

[0096] 3) AA weighted common adjacent node weight (Adamic / Adar, AA)

[0097] For the enterprise node pair s u , s v 's common adjacent node s k ∈ Ne i (s u ) I Ne i (s v ), the influence of different adjacent nodes on the reuse probability of these two enterprise nodes is different. For example, for an adjacent node s k If there are many adjacent nodes, then its influence on these two enterprise node pairs will be significantly reduced. The Adamic-Adar weighting method is used to weight the effect of common adjacent nodes, and the following AA weighted common adjacent node weight is defined:

[0098]

[0099] 4) Reciprocal Last Edge Occurred TimeStamps (RLE)

[0100] In the industrial collaboration network, due to the continuous addition of new enterprises and the emergence of new ecological communities, leading enterprises pay more attention to fresh ecological community organization models. If an organization model has not been used for a long time, then the probability of its being reused will decrease significantly; therefore, the following index is defined from the perspective of the time interval since the last occurrence of the node connection edge:

[0101]

[0102] LE i (s u , s v ) represents the time interval since the last occurrence of the node connection edge. Obviously, the larger the time interval, the smaller the probability of the node connection edge appearing again. Therefore, its reciprocal form is defined as an index for evolutionary prediction:

[0103]

[0104] 2. Node feature definition

[0105] Node characteristics represent judging the probability of a node reappearing in the future from the perspective of the node. If the probabilities of two nodes reappearing in the future are relatively high, then these two nodes will form a node connection edge with a relatively high probability in the future;

[0106] 1) Vertex Weighted Degree (VWD) of enterprise nodes

[0107] The cumulative distribution of the weighted degree of enterprise nodes satisfies the power-law characteristic, that is, the higher the weighted degree of the node, the higher the probability of being reused in the future. This is mainly because enterprise nodes with a high weighted degree often participate in industrial clusters, and leading enterprises are more likely to trust such enterprises. Therefore, the following evolutionary characteristics are defined from the perspective of the weighted degree of nodes:

[0108]

[0109] 2) Page Rank based Service Importance (PRI) of nodes

[0110] The PageRank algorithm is widely used in networks to evaluate the importance of each node in the network. In the industrial collaboration network, if the adjacent nodes of a node have a high probability of being reused, and since this node can form a collaboration relationship with its adjacent nodes to meet certain functions, then this node will also have a relatively high probability of being concerned by leading enterprises; based on the importance of PageRank, the following evolutionary characteristics are defined:

[0111] PRI i (s u ,s v ) = PRI i (s u ) + PRI i (s v )

[0112] where PRI i (s u ) represents the PageRank importance of enterprise node s u in the collaboration network;

[0113] 3) Betweenness based Service Centrality (BSC) of nodes

[0114] In the collaboration network, if a node can establish connections with other nodes at a shorter distance, then this node will have a higher probability of being selected by leading enterprises in the future; the Betwenness betweenness centrality index is used to reflect the distance between network nodes and other nodes, and a higher Betweenness centrality indicates that the node has a higher centrality:

[0115] BSC i (su , s v ) = BSC i (s u ) + BSC i (s v )

[0116] Among them, BSC i (s u ) represents the Betweenness centrality of enterprise node s u in the collaboration network;

[0117] 4) The time interval when the enterprise node was last used (Reciprocal Last Vertex Occurred TimeStamps, RLV)

[0118] If a node has not been used for a long time, the probability of the node being used again will continuously decrease. Therefore, the following last node usage time interval LV i (s u ) is defined as:

[0119]

[0120] Obviously, the larger the time interval, the smaller the probability of the node reappearing. Therefore, the reciprocal form is used to define the probability of two enterprise nodes reappearing:

[0121]

[0122] As can be seen from the above definitions, the six indicators of CSEW, WCN, AA, VWD, PRI, and BSC are mainly defined from the node pairs and the topological structure of the nodes in a single industrial collaboration network, while the two indicators of RLE and RLV involve the time characteristics of the node pairs and the nodes in multiple industrial collaboration networks. Therefore, these eight indicators can be organized into a Figure 3 two-dimensional grid as shown.

[0123] Thus, the embodiments of the present application define eight evolution characteristics of enterprise node pairs in the industrial collaboration network from two perspectives of enterprise nodes and enterprise node pairs, network topological structure and time characteristics, form a two-dimensional industrial collaboration network evolution characteristic set, and then can analyze the ability of different evolution characteristics to reflect the evolution process of the industrial collaboration network.

[0124] In step S103, according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction, the evolution process of the industrial Internet ecosystem is predicted.

[0125] After extracting the evolutionary characteristics of the industrial collaboration network, the embodiments of the present application can perform predictive analysis on the evolutionary process of the industrial collaboration network from the perspective of network prediction, and give a recommendation method based on the predicted network.

[0126] Optionally, in an embodiment of the present application, predicting the industrial Internet ecological evolution process according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction includes: based on a preset unsupervised linear classifier, analyzing the reflection degree of at least one index on the evolutionary characteristics of the industrial collaboration network to the evolutionary process of the industrial collaboration network; quantifying the reflection degree, and using the ranking of the characteristic values corresponding to the nodes as the index for node evolution prediction.

[0127] After extracting the evolutionary characteristics of the industrial collaboration network, the embodiments of the present application can verify the evolutionary characteristics of the industrial collaboration network, and the specific process is as follows:

[0128] 1. Definition of unsupervised linear classifier:

[0129] In order to analyze the reflection degree of the eight indexes defined above on the evolutionary characteristics of the industrial collaboration network to the evolutionary process of the industrial collaboration network, the embodiments of the present application need to define an unsupervised linear classifier in the following way:

[0130] Given a network evolution feature f and a threshold fv, for any node pair s in the network u , s v , calculate its eigenvalue f(s u , s v ). If f(s u , s v ) ≥ fv, it is considered that this node pair will build an enterprise collaboration edge in the future; otherwise, if f(s u , s v ) < fv, it is considered that this node will not form an edge in the future;

[0131] 2. Quantitative analysis of prediction results:

[0132] In order to quantify and define the prediction effect of this classifier on the network evolution process, it is necessary to verify the prediction result according to the actual state of the node pair in the network, and use the following Recall / Fallout indicators to quantitatively analyze the prediction result:

[0133]

[0134] Among them, true positives represent the number of node pairs that the linear classifier believes will appear in and actually appear; false positives represent the number of node pairs that the classifier believes will appear in The number of node pairs that should have appeared but did not; positives represents the number of enterprise node pairs that appeared in , negatives represents the number of node pairs that did not appear in ;

[0135] From the above definitions, it can be obtained that Recall / Fallout represents the multiple of the probability of a node pair appearing in the future network relative to the probability of not appearing; obviously, the higher the index value, the better the classification effect of the classifier. For example, if the Recall / Fallout value of a node pair's CSEW reaches 100, it means that the probability of this node appearing in the future collaboration network will be 100 times the probability of not appearing;

[0136] Therefore, given a collaboration network and a feature f, the range of values of this feature corresponding to all nodes in it can be obtained; by varying the threshold fv within the range of values, the process of the manifestation of this feature on the evolution effect of this collaboration network can be obtained;

[0137] Figure 4 shows, taking as an example, the prediction effects of each feature value on . It can be seen from the figure that for all eight feature values, as the threshold increases, their Recall / Fallout index values continuously increase, which indicates that for the node pairs in the network , the higher the feature value, the higher the probability of appearing in the future network , and the node pairs that appear in the network are relatively concentrated in the area with relatively high feature values. Therefore, the ranking of these feature values can be used as an indicator for the evolution prediction of node pairs;

[0138] 3. Normalized summary of prediction results

[0139] From the above analysis, it can be seen that each feature can reflect the evolution law of the collaboration network, and the larger the feature value, the higher the probability of this node pair forming an industrial ecological community in the future; therefore, the feature values of all node pairs can be sorted, and the higher the ranking, the higher the probability of appearance; however, each feature only reflects one aspect of the evolution law of the collaboration network. Therefore, the embodiments of this application can use the method of sorting fusion to achieve the summary of all feature values, and the specific process is as follows:

[0140] 1) First, assume that represents a node pair (s in the collaboration network u , s v), which can be simplified to Here j = 1, KN, Representing collaborative networks All possible node pairs; for each feature f, calculate the eigenvalue of each node pair and sort them from large to small according to their eigenvalues to obtain the basic ranking sequence (Primary Ranking): f={CSEW,WCN,AA,RLE,VWD,PRI,BSC,RLV};

[0141] 2) Since different sorting sequences are calculated based on different eigenvalues, the range of their absolute values is very different. For example, the range of CSEW is [1, 22], while the range of RLE is [0.1, 1]. Therefore, it is meaningless to directly use the eigenvalue. In addition, due to the obvious power law characteristics in the industrial collaboration network, there is a significant absolute value gap in the values in the sorting sequence. In addition, since the ultimate goal is to sort all node pairs, the position of the node pair in the sorting sequence has a clearer meaning. Therefore, according to the Borda scoring method, the node pairs are re-normalized:

[0142]

[0143] in, Represents a node pair In the basic sort sequence The position among them, obviously for the node pairs with higher ranking, The smaller the value of is, the higher the Borda value is;

[0144] 3) After the above Borda normalization, the difference in domain values between different sorting sequences is effectively eliminated, and the unevenness problem caused by the power law characteristics is also eliminated. Therefore, the embodiment of the present application can directly use a linear function to linearly add the Borda score values of each node in each sorting sequence:

[0145] F=RW

[0146] in, W=[w1,K w8] T , the embodiments of the present application can be directly adopted It represents the impact of each eigenvalue on the evolution of the industrial collaboration network.

[0147] It is understandable that the embodiments of the present application can start from the industrial ecosystem collaboration network, define the sharded industrial collaboration network, the past industrial collaboration network, and the forward industrial collaboration network, form an industrial collaboration network sequence model, define the evolutionary prediction problem of the industrial collaboration network from the perspective of network sequence evolution, give a formal definition of the problem from the perspective of link prediction; define eight evolutionary characteristics of enterprise node pairs in the industrial collaboration network from two perspectives of enterprise nodes and enterprise node pairs, network topology and time characteristics, form a two-dimensional industrial collaboration network evolutionary characteristic set, and analyze the ability of different evolutionary characteristics to reflect the evolutionary process of the industrial collaboration network, and verify the effectiveness of the eight evolutionary characteristics; and then propose a sorting fusion method to fuse the above evolutionary characteristics to predict the evolutionary process of the industrial collaboration network.

[0148] The industrial Internet ecological evolution prediction method based on link prediction proposed according to the embodiments of the present application obtains an industrial collaboration network sequence model and a formal definition of the problem; then conducts a prediction analysis on the evolution of the industrial collaboration network from the perspective of network prediction; and finally gives a recommendation method based on the predicted network, so as to effectively predict the evolution of the industrial collaboration network according to the historical industrial collaboration network, and extract successful ecological community organization methods from it to support the recommendation of cooperative enterprises.

[0149] Next, a description is given with reference to the drawings of an industrial Internet ecological evolution prediction device based on link prediction proposed according to the embodiments of the present application.

[0150] Figure 5 It is a block diagram of an industrial Internet ecological evolution prediction device based on link prediction according to the embodiments of the present application.

[0151] As Figure 5 shown, the industrial Internet ecological evolution prediction device 10 based on link prediction includes: a definition module 100, a verification module 200, and a prediction module 300.

[0152] Among them, the definition module 100 is used to determine an industrial collaboration network sequence model and a formal definition of the problem.

[0153] The verification module 200 is used to extract at least one industrial collaboration network evolutionary characteristic based on the industrial collaboration network sequence model and the formal definition of the problem, and verify at least one industrial collaboration network evolutionary characteristic to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction.

[0154] The prediction module 300 is used to predict the industrial Internet ecological evolution process according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction.

[0155] Optionally, in an embodiment of the present application, the definition module includes: a first generation unit, a processing unit, a second generation unit, an acquisition unit, and a modeling unit.

[0156] Among them, the first generation unit is used to construct a collaboration relationship based on the organization mode of the ecological community formed within the target time period, so as to generate a fragmented industry collaboration network.

[0157] The processing unit is used to obtain a fragmented industry collaboration network sequence from the fragmented industry collaboration networks of each time slice of the target time series.

[0158] The second generation unit is used to generate a past industry collaboration network and a forward industry collaboration network based on the fragmented industry collaboration network sequence.

[0159] The acquisition unit is used to obtain a past industry collaboration matrix and a forward industry collaboration matrix according to the past industry collaboration network and the forward industry collaboration network.

[0160] The modeling unit is used to predict the future network of the fragmented industry collaboration network sequence, and based on the past industry collaboration network and the forward industry collaboration network, while forming an industry collaboration network sequence model, determine the formal definition of the problem.

[0161] Optionally, in an embodiment of the present application, the past industry collaboration matrix is expressed as:

[0162]

[0163] Among them, t i is the i-th time slice, l is the time window, θ is the time decay parameter, represents the fragmented industry collaboration matrix;

[0164] And, the forward industry collaboration matrix is expressed as:

[0165]

[0166] Among them, g is the forward time window.

[0167] Optionally, in an embodiment of the present application, the modeling unit is specifically used to, based on the given fragmented industry collaboration network sequence, past time window, time decay parameter, and forward time window, construct the corresponding past industry collaboration network and the corresponding forward industry collaboration network, and find a mapping function to calculate the probability of occurrence in the corresponding forward industry collaboration network for any pair of node pairs in the corresponding past industry collaboration network.

[0168] Optionally, in an embodiment of the present application, the prediction module 300 includes: an analysis unit and a quantization unit.

[0169] Among them, an analysis unit is configured to analyze, based on a preset unsupervised linear classifier, the degree to which at least one metric reflects the evolutionary characteristics of the industrial collaboration network in the process of the evolution of the industrial collaboration network.

[0170] A quantization unit is configured to quantify the degree of reflection, and use the sorting of the eigenvalue corresponding to the node as an index for node evolution prediction.

[0171] It should be noted that the foregoing explanation of the embodiment of the method for predicting the evolution of the industrial Internet ecosystem based on link prediction also applies to the device for predicting the evolution of the industrial Internet ecosystem based on link prediction in this embodiment, and will not be elaborated here.

[0172] According to the device for predicting the evolution of the industrial Internet ecosystem based on link prediction proposed in the embodiment of the present application, by determining the industrial collaboration network sequence model and the formal definition of the problem; based on the industrial collaboration network sequence model and the formal definition of the problem, extracting at least one evolutionary characteristic of the industrial collaboration network, and verifying at least one evolutionary characteristic of the industrial collaboration network to obtain a prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; predicting the evolution process of the industrial Internet ecosystem according to the prediction analysis result of the evolution of the industrial collaboration network from the perspective of network prediction, so as to effectively predict the evolution of the industrial collaboration network according to the historical industrial collaboration network, and extract successful ecological community organization methods therefrom to support the recommendation of cooperative enterprises.

[0173] Figure 6 The figure is a schematic structural diagram of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0174] A memory 601, a processor 602, and a computer program stored on the memory 601 and executable on the processor 602.

[0175] When the processor 602 executes the program, it implements the method for predicting the evolution of the industrial Internet ecosystem based on link prediction provided in the foregoing embodiment.

[0176] Furthermore, the electronic device further includes:

[0177] A communication interface 603, configured for communication between the memory 601 and the processor 602.

[0178] The memory 601 is used for storing a computer program executable on the processor 602.

[0179] The memory 601 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.

[0180] If the memory 601, the processor 602, and the communication interface 603 are implemented independently, the communication interface 603, the memory 601, and the processor 602 can be interconnected via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 it is represented by only a thick line in Figure 6 , but this does not mean that there is only one bus or one type of bus.

[0181] Optionally, in a specific implementation, if the memory 601, the processor 602, and the communication interface 603 are integrated on a single chip, the memory 601, the processor 602, and the communication interface 603 can communicate with each other through an internal interface.

[0182] The processor 602 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0183] The embodiments of the present application also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned method for predicting the evolution of the industrial Internet ecosystem based on link prediction is implemented.

[0184] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0185] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the technical features indicated. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0186] Any process or method description represented in a flowchart or described otherwise herein can be understood to represent a module, segment, or portion of code including one or N executable instructions for implementing a customized logical function or process. The scope of the preferred embodiments of the present application includes additional implementations, where functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present application pertain.

[0187] The logic and / or steps represented in a flowchart or described otherwise herein, for example, can be considered a sequenced list of executable instructions for implementing a logical function, and can be embodied specifically in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion having one or N wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.

[0188] It should be understood that each part of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one of the following techniques known in the art or a combination thereof can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0189] Those of ordinary skill in the art can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0190] In addition, in each embodiment of the present application, the functional units can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into a module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. When the above integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0191] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disc, or the like. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. An industrial Internet ecological evolution prediction method based on link prediction, characterized in that Including the following steps: Determine the industrial collaboration network sequence model and the formal definition of the problem; Based on the industrial collaboration network sequence model and the formal definition of the problem, extract at least one industrial collaboration network evolution feature, and verify the at least one industrial collaboration network evolution feature to obtain a predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; And Predict the evolution process of the industrial Internet ecosystem according to the predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; The determination of the industrial collaboration network sequence model and the formal definition of the problem includes: Construct collaboration relationships based on the organizational mode of the ecological community formed within the target time period to generate a fragmented industrial collaboration network; Obtain a fragmented industrial collaboration network sequence from the fragmented industrial collaboration networks of each time slice of the target time series; Generate a past industrial collaboration network and a forward industrial collaboration network based on the fragmented industrial collaboration network sequence; Obtain a past industrial collaboration matrix and a forward industrial collaboration matrix according to the past industrial collaboration network and the forward industrial collaboration network; Predict the network of the fragmented industrial collaboration network sequence in the future, and based on the past industrial collaboration network and the forward industrial collaboration network, while forming the industrial collaboration network sequence model, determine the formal definition of the problem; The past industrial collaboration matrix is expressed as: where t i is the i-th time slice, l is the time window, and θ is the time decay parameter, represents the slice industry collaboration matrix; And the forward industrial collaboration matrix is expressed as: Where g is the forward time window.

2. The method according to claim 1, wherein The determination of the formal definition of the problem includes: Based on the given fragmented industrial collaboration network sequence, past time window, time decay parameter, and forward time window, construct the corresponding past industrial collaboration network and the corresponding forward industrial collaboration network, and find a mapping function to calculate the probability of occurrence in the corresponding forward industrial collaboration network for any pair of nodes in the corresponding past industrial collaboration network.

3. The method according to claim 1, characterized in that, The prediction of the evolution process of the industrial Internet ecosystem according to the predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction includes: Based on a preset unsupervised linear classifier, analyze the degree to which at least one indicator reflects the industrial collaboration network evolution characteristics on the industrial collaboration network evolution process; Quantify the degree of reflection, and use the ranking of the characteristic values corresponding to the nodes as an indicator for the evolution prediction of the nodes.

4. An industrial Internet ecological evolution prediction device based on link prediction, characterized in that, Including: A definition module for determining the industrial collaboration network sequence model and the formal definition of the problem; A verification module for extracting at least one industrial collaboration network evolution feature based on the industrial collaboration network sequence model and the formal definition of the problem, and verifying the at least one industrial collaboration network evolution feature to obtain a predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; And A prediction module for predicting the evolution process of the industrial Internet ecosystem according to the predictive analysis result of the evolution of the industrial collaboration network from the perspective of network prediction; The definition module includes: A first generation unit for constructing collaboration relationships based on the organizational mode of the ecological community formed within the target time period to generate a fragmented industrial collaboration network; A processing unit for obtaining a fragmented industrial collaboration network sequence from the fragmented industrial collaboration networks of each time slice of the target time series; A second generation unit for generating a past industrial collaboration network and a forward industrial collaboration network based on the fragmented industrial collaboration network sequence; An acquisition unit for obtaining a past industrial collaboration matrix and a forward industrial collaboration matrix according to the past industrial collaboration network and the forward industrial collaboration network; A modeling unit for predicting the network of the fragmented industrial collaboration network sequence in the future, and forming the industrial collaboration network sequence model based on the past industrial collaboration network and the forward industrial collaboration network, while determining the formal definition of the problem; The past industrial collaboration matrix is expressed as: where t i is the i-th time slice, l is the time window, and θ is the time decay parameter, represents the sliced industry collaboration matrix; And the forward industrial collaboration matrix is expressed as: Where g is the forward time window.

5. The device according to claim 4, characterized in that The modeling unit is specifically used for Based on the given fragmented industrial collaboration network sequence, past time window, time decay parameter, and forward time window, constructing the corresponding past industrial collaboration network and the corresponding forward industrial collaboration network, and finding a mapping function to calculate the probability of occurrence in the corresponding forward industrial collaboration network for any pair of nodes in the corresponding past industrial collaboration network.

6. The device according to claim 4, characterized in that, The prediction module includes: An analysis unit for analyzing the degree to which at least one indicator reflects the evolution characteristics of the industrial collaboration network in the industrial collaboration network evolution process based on a preset unsupervised linear classifier; A quantification unit for quantifying the reflection degree, and using the ranking of the eigenvalue corresponding to the node as an indicator for the evolution prediction of the node pair.

7. An electronic device, characterized in that, Including: A memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the industrial Internet ecological evolution prediction method based on link prediction according to any one of claims 1-3.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to be used to implement the industrial Internet ecological evolution prediction method based on link prediction according to any one of claims 1-3.

Images