Futures network-based investment risk control method

Abstract

The invention relates to the technical field of investment risk control, in particular to an investment risk control method based on a futures network, and the method comprises the steps: obtaining historical price data in futures transaction; entropy weight fusion is carried out on the price data of the futures transaction to obtain fusion data; the fusion data is visualized, and a futures network is constructed; calculating a centrality index of each node in the futures network; fusing the centrality indexes of each node, and sorting the nodes based on a fusion result to obtain key risk nodes and non-key risk nodes; and regulating and controlling the futures varieties corresponding to the key risk nodes or / and the non-key risk nodes. According to the method, the dependence of a traditional risk assessment method on a single financial index is overcome, the internal association relationship of the futures market can be more deeply mined, and a scientific risk control basis and optimization strategy support are provided for investors.

Description

Technical Field

[0001] The present invention relates to the technical field of investment risk control, and in particular to an investment risk control method based on a futures network. Background Art

[0002] With the continuous development and changes of the global economy, the volatility of financial markets, especially futures markets, is increasing. In this environment, investors are facing more complex market risks, especially in futures markets such as exchange rates, commodity prices, and stock indexes. How to effectively identify and control investment risks has become a key issue in investment decision-making.

[0003] Traditional risk assessment methods usually rely on a single financial indicator or a statistical model based on historical data. However, these methods often ignore the complex correlations and nonlinear characteristics in the futures market, and cannot fully reflect the mutual influence and potential risks between different market factors. In recent years, complex network theory has been widely used in the study of financial markets, especially in futures trading networks. Complex network theory can reveal the correlation between different futures products and the potential systemic risks in the market. By constructing a futures trading network model, the mutual influence between different futures products in the market can be quantitatively analyzed, thereby identifying key risk nodes and their potential impacts.

[0004] However, most existing risk assessment methods focus on the application of a single centrality indicator and fail to fully consider the advantages of multi-dimensional information fusion, which limits the accuracy and effectiveness of risk assessment. In order to more effectively deal with the complexity and uncertainty in the market, a comprehensive risk assessment and control method is urgently needed, which can identify and evaluate the key risk nodes in futures trading through scientific algorithms under the background of multiple factors and multiple dimensions, so as to provide investors with more accurate risk control strategies. Summary of the invention

[0005] The purpose of the present invention is to provide an investment risk control method based on futures networks, which combines complex network theory with evidence theory and proposes a comprehensive risk assessment and control strategy based on multiple centrality indicators, thereby overcoming the limitation of traditional methods that rely on a single indicator.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows: An investment risk control method based on futures network comprises the following steps: S1. Obtain historical price data in futures trading; S2, perform entropy weight fusion on the price data of futures transactions to obtain fused data; S3, visualize the fused data and build a futures network; S4. Calculate the centrality index of each node in the futures network; S5. Fusion the centrality index of each node, sort the nodes based on the fusion result, and obtain key risk nodes and non-key risk nodes; S6. Regulate the futures products corresponding to key risk nodes and / or non-key risk nodes.

[0007] Furthermore, the price data of each type of futures transaction is expressed as ,in is the number of consecutive trading days, It is the trading data of each trading day, including the opening price, closing price, highest price and lowest price. Each type of trading data under consecutive trading days constitutes a time series.

[0008] Furthermore, the entropy weight fusion of the price data of futures transactions includes the following steps: S21. Calculate the scale matrix : ; For price data An element in indicates that the t-th trading day is d The value of the class transaction data, t∈T, d ∈D; For price data No. The sum of the column elements; S22. Calculate logarithm : ; S23. Calculating Entropy : ; is the normalization coefficient; S24. Calculate weight : ; S25, by weight The transaction data of the four time series of opening price, closing price, highest price and lowest price are merged into one time series of fused data.

[0009] Furthermore, the S3 includes: S31, taking each fused data in the time series as a node in the futures network; S32, for each fused data in the time series, judging whether there is a connection between two fused data through a visualization formula; S33, if there is a connection between the two fused data, establishing an edge between the two fused data; S34, repeating S32~S33 to construct a futures network.

[0010] Furthermore, the visualization formula is as follows: ; Set the number of consecutive trading days T is the time variable, , the fused data set under the time variable is recorded as X , ， ,b,c∈n, for and Any point between Satisfies the visualization formula, where , indicating the fusion data set X Fusion Data and x b A connection exists.

[0011] Furthermore, the centrality indexes include degree value, closeness centrality, harmonic closure centrality, betweenness centrality, PageRank value, and eigenvector centrality. i≤n ; The degree value is calculated as follows: ; Is a node i The degree value, Is a node i and nodes j If the edges between nodes i and nodes j If there is an edge between The value of is defined as 1, otherwise it is 0; The closeness centrality is calculated as follows:

[0012] Is a node i The closeness centrality, n is the number of nodes in the futures network, Is a node i To Node j The shortest path length; The harmonic closure centrality is calculated as follows: ; Is a node i The harmonic closure centrality of n is the number of nodes in the futures network, Is a node i To Node j The shortest path length; The betweenness centrality is calculated as follows: ; Is a node i The betweenness centrality of , v, s, t all represent nodes, It is a slave node s To Node t The number of shortest paths, It is a slave node s To Node t The shortest path passes through the node i The number of shortest paths; The PageRank value is calculated as follows: ; Is a node i In the t The PageRank value of the iteration, Is a node i In the t-1 The PageRank value of the iteration, is an indicator function, if the node j To Node i If there is a direct connection, = 1, otherwise = 0, For Node j The out-degree of The eigenvector centrality is calculated as follows: ; Is a node i The eigenvector centrality of is the adjacency matrix of the futures network, λ is the largest eigenvalue of the adjacency matrix, Is a node i The corresponding feature vector.

[0013] Furthermore, in S5, the centrality index of each node is fused based on the DS evidence theory to obtain the importance characterization value of each node, and the nodes are sorted from large to small according to the importance characterization value. The top N nodes in the sorting are recorded as key risk nodes, and the remaining nodes are recorded as non-key risk nodes.

[0014] Furthermore, based on the DS evidence theory, the normalized values ​​of the centrality index of each node are fused to obtain the importance characterization value of each node, including the following operations: S51, calculate the normalization coefficient 1- K : ; S52, according to the normalization coefficient 1- K Calculate the importance of each node ; .

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention collects historical price data in futures transactions, generates a futures network using a time series-based visualization algorithm, calculates the centrality index of each node in the network to quantify the importance of the node, and combines multiple centrality indexes with evidence theory for fusion analysis, thereby accurately identifying key risk nodes in the futures network. Based on the identification results of key risk nodes, the potential risks of futures transactions are comprehensively evaluated and effectively controlled.

[0016] The present invention overcomes the reliance of traditional risk assessment methods on a single financial indicator, and can more deeply explore the inherent correlations in the futures market, providing investors with a scientific basis for risk control and support for optimization strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of the method of the present invention.

[0018] Figure 2 A schematic diagram for visualization.

[0019] Figure 3 A network diagram generated using visualization. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the present invention provides a futures network-based investment risk control method, comprising the following steps: S1. Obtain historical price data in futures trading; S2, perform entropy weight fusion on the price data of futures transactions to obtain fused data; S3, visualize the fused data and build a futures network; S4. Calculate the centrality index of each node in the futures network; S5. Fusion the centrality index of each node, sort the nodes based on the fusion result, and obtain key risk nodes and non-key risk nodes; S6. Regulate the futures products corresponding to key risk nodes and / or non-key risk nodes.

[0022] The present invention constructs a futures network based on historical transaction data in futures trading, and identifies and assesses key risk nodes through the centrality index of nodes in the complex network. Different from the traditional risk assessment method that relies on a single financial indicator, the present invention identifies key risk nodes with significant influence by analyzing the inherent correlation of the futures market, thereby providing investors with accurate risk assessment and optimized investment decision support.

[0023] First, obtain the historical price data of futures transactions. The price data of each type of futures transaction is expressed as ,in is the number of consecutive trading days, It is the trading data of each trading day, including the opening price, closing price, highest price and lowest price. Each type of trading data under consecutive trading days constitutes a time series.

[0024] Then the fused data is visualized to construct a futures network. Specifically, the entropy weight method is used to perform entropy weight fusion on transaction data of different dimensions (i.e., opening price, closing price, highest price, and lowest price), including the following steps: S21. Calculate the scale matrix : ; For price data An element in indicates that the t-th trading day is d The value of the class transaction data, t∈T, d ∈D; For price data No. The sum of the column elements; S22. Calculate logarithm : ; S23. Calculating Entropy : ; is the normalization coefficient; S24. Calculate weight : ; in represents the information utility value, It is weight normalization processing; S25, by weight The transaction data of the four time series of opening price, closing price, highest price and lowest price are merged into one time series of fused data.

[0025] Then, the fused data is visualized to construct a futures network. Specifically, it includes: S31, taking each fused data in the time series as a node in the futures network; S32, for each fused data in the time series, judging whether two fused data are connected by a visualization formula; S33, if two fused data are connected, an edge is established between the two fused data; S34, repeating S32~S33 to construct a futures network.

[0026] The principle of visualization is as follows Figure 2 As shown, the fused data in the time series corresponds to the nodes in the futures network, and the lines between the fused data that satisfy the visualization formula are defined as the edges of the network, that is, if the tops of two cuboids are visible to each other, then a straight line is used to connect the two tops. The visualization formula is as follows: .

[0027] Set the number of consecutive trading days T is the time variable, , the fused data set under the time variable is recorded as X , ， ,b,c∈n, for and Any point between Satisfies the visualization formula, where , indicating the fusion data set X Fusion Data and x b A connection exists.

[0028] The network diagram generated by visualization is as follows Figure 3 As shown, the solid points in the figure correspond to the data points of the time series one by one, and the lines between the solid points correspond to the lines between the cuboids one by one. Through the visualization of the futures network constructed, each node in the futures network corresponds to a fusion data in the time series.

[0029] Based on the futures network, the centrality index of each node in the futures network is calculated. The centrality index includes degree value, closeness centrality, harmonic closure centrality, betweenness centrality, PageRank value, eigenvector centrality, i≤n ; The degree is used to quantify the number of connections of an individual in the network, that is, the degree of a node is the number of other nodes directly connected to it. The degree is calculated as follows: Set the number of consecutive trading days T Time variable , the fused data under the time variable is recorded as X , ,if and Two points are visible, then for any point ,in, Satisfies the visualization formula.

[0030] Furthermore, the centrality indexes include degree value, closeness centrality, harmonic closure centrality, betweenness centrality, PageRank value, and eigenvector centrality. i≤n ; The degree value is calculated as follows: ; Is a node i The degree value, Is a node i and nodes j If the edges between nodes i and nodes j If there is an edge between The value of is defined as 1 if it is set, and 0 otherwise.

[0031] For undirected graphs, closeness centrality represents the inverse of the average shortest path length from a node to other nodes. Closeness centrality is calculated as follows: ; Is a node i The closeness centrality, n is the number of nodes in the futures network, Is a node i To Node j The shortest path length.

[0032] Harmonic closure centrality represents the harmonic mean of the shortest path lengths from a node to other nodes. Harmonic closure centrality is calculated as follows: ; Is a node i The harmonic closure centrality of n is the number of nodes in the futures network, Is a node i To Node j The shortest path length.

[0033] Betweenness centrality is used to measure the degree to which a node's position in a network controls the communication flow between other nodes in the network. The calculation of betweenness centrality is as follows: ; Is a node i The betweenness centrality of , v, s, t all represent nodes, It is a slave node s To Node t The number of shortest paths, It is a slave node s To Node t The shortest path passes through the node i The number of shortest paths.

[0034] The PageRank value is an indicator to measure the nodes in the network. The higher the PageRank value of a node, the greater its influence in the network. The PageRank value is calculated as follows: ; Is a node i In the t The PageRank value of the iteration, Is a node i In the t-1 The PageRank value of the iteration, is an indicator function, if the node j To Node i If there is a direct connection, = 1, otherwise a ij = 0, For Node j The out-degree of .

[0035] Eigenvector centrality is used to measure the importance of nodes in the network. The larger the value of eigenvector centrality, the greater the influence of the node in the network. The calculation of eigenvector centrality is as follows: ; Is a node i The eigenvector centrality of is the adjacency matrix of the futures network, λ is the largest eigenvalue of the adjacency matrix, Is a node i The corresponding feature vector.

[0036] Then, based on the DS evidence theory, the centrality index of each node is fused to obtain the importance characterization value of each node. The nodes are sorted from large to small according to the importance characterization value. The top N nodes in the sorting are recorded as key risk nodes, and the remaining nodes are recorded as non-key risk nodes.

[0037] DS Evidence Theory is an existing mathematical framework for dealing with uncertainty and reasoning. It is mainly used for reasoning under uncertainty or missing information, rather than relying on traditional probability theory. The basic idea of ​​DS Evidence Theory is to assign confidence to different hypotheses, allowing for more flexible modeling of uncertainty.

[0038] The present invention fuses the normalized value of the centrality index of each node based on the DS evidence theory to obtain the importance characterization value of each node, including the following operations: S51, calculate the normalization coefficient 1- K : ; S52, according to the normalization coefficient 1- K Calculate the importance of each node ; .

[0039] The key risk nodes identified by the present invention are located at the peaks or troughs of the time series, and the peaks or troughs indicate changes in price trends, that is, the market factors corresponding to that moment affect the price trend.

[0040] Finally, the futures products corresponding to the key risk nodes and / or non-key risk nodes are regulated. The regulation includes monitoring dynamic risks, optimizing investment portfolios, managing weighted risk exposures, and adopting risk mitigation strategies.

[0041] Monitoring dynamic risks includes: (1) Investors will first track and monitor the key risk nodes in the futures market. These nodes are the result of historical market data and complex network analysis, reflecting the risk characteristics of futures products. The higher the importance of each node (such as the centrality index), the greater the impact of the futures product on market fluctuations and the greater the risk. For key risk nodes, investors should regularly evaluate their importance in the futures network and adjust their futures positions accordingly.

[0042] (2) For key risk nodes at the peak of risk, investors should adjust futures positions in a timely manner according to the importance of the node to reduce the risk of over-concentrated investment. For example, when the risk value of a certain node is high, that is, the node is a key risk node, investors can reduce the position of the futures product according to the importance of the node to avoid over-concentrated investment. On the contrary, if the risk value of the node is low, that is, the node is a non-key risk node, investors can increase the investment exposure of the futures product to balance the overall risk of the investment portfolio.

[0043] Optimizing your portfolio includes: (1) According to the ranking of node importance, investors can adjust their investment portfolios according to risk control needs. Specifically, they should reduce investment in futures products that are near the peak of risk volatility. In other words, investors can adjust the composition of their futures portfolios according to the ranking of nodes. Futures products at key risk nodes often show greater market volatility and higher risks, so their weight in the investment portfolio needs to be reduced.

[0044] (2) For key risk nodes at the peak of risk, investors can reduce the investment proportion of the corresponding futures products. For example, when the node of a futures product is close to the peak of risk fluctuation, it indicates that its market risk is relatively high, and investors should consider reducing their positions. On the contrary, for non-key risk nodes with lower risk values, their market performance is relatively stable and is not easily affected by large external fluctuations. Investors can moderately increase the investment proportion of the corresponding futures products and use their relatively stable market trends to diversify risks.

[0045] Managing weighted risk exposures includes: (1) Based on the importance characterization value, different investment exposures are set for the futures products corresponding to the nodes. The larger the importance characterization value, the greater the risk faced by the futures product in the market, so the exposure of these futures products should be reduced; while the lower the importance characterization value, the larger the exposure can be set and the investment can be increased.

[0046] (2) Adjust exposure regularly. As the market environment and nodes change, investors need to regularly adjust the investment exposure of each futures product. For example, when the risk of a certain node begins to rise significantly, investors should reduce their exposure to that futures product; and when the risk of a certain node decreases, they can increase their investment exposure accordingly.

[0047] Risk mitigation strategies include: (1) When the market fluctuates violently or the node risk of futures products increases significantly, that is, at the key risk node where the risk peaks, investors should adopt risk mitigation strategies such as hedging and stop loss to reduce potential losses. Different mitigation measures can be taken based on the importance of the node.

[0048] (2) For key risk nodes at the peak of risk, investors can adopt futures hedging strategies, such as option hedging or futures long-short hedging strategies at key risk nodes. Through hedging operations, investors can reduce the risk exposure of futures products and reduce losses caused by market fluctuations.

[0049] (3) When the market fluctuates violently, investors should promptly implement stop-loss strategies based on the risk signals of the nodes. For example, when the price fluctuation of a certain node is too large and the importance characterization value increases, investors can automatically trigger stop-loss measures to prevent further losses. Investors can also combine the preset risk warning system to obtain node risk data in real time and respond quickly.

[0050] Finally, it should be noted that the above embodiments are only preferred embodiments of the present invention to illustrate the technical solutions of the present invention, rather than limiting them, and certainly not limiting the patent scope of the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention. In other words, any changes or modifications made to the main design concept and spirit of the present invention that have no substantive significance, and the technical problems they solve are still consistent with the present invention, should be included in the protection scope of the present invention. In addition, the direct or indirect application of the technical solutions of the present invention in other related technical fields is also included in the patent protection scope of the present invention.

Claims

1. A method for controlling investment risks based on futures network, characterized in that: The following steps are involved: S1. Obtain historical price data in futures trading; S2, perform entropy weight fusion on the price data of futures transactions to obtain fused data; S3, visualize the fused data and build a futures network; S4, calculate the centrality index of each node in the futures network; S5. Fusion the centrality index of each node, sort the nodes based on the fusion result, and obtain key risk nodes and non-key risk nodes; S6. Regulate the futures products corresponding to key risk nodes and / or non-key risk nodes.

2. The investment risk control method based on futures network according to claim 1, characterized in that: The price data for each type of futures transaction is expressed as ,in is the number of consecutive trading days, It is the trading data of each trading day, including the opening price, closing price, highest price and lowest price. Each type of trading data under consecutive trading days constitutes a time series.

3. The investment risk control method based on futures network according to claim 2 is characterized in that: The entropy weight fusion of the price data of futures transactions comprises the following steps: S21. Calculate the scale matrix : ； For price data An element in indicates that the t-th trading day is d The value of the class transaction data, t∈T, d ∈D; For price data No. The sum of the column elements; S22. Calculate logarithm : ； S23. Calculating Entropy : ； is the normalization coefficient; S24. Calculate weight : ； S25, by weight The transaction data of the four time series of opening price, closing price, highest price and lowest price are merged into one time series of fused data.

4. The investment risk control method based on futures network according to claim 1 is characterized in that: The S3 includes: S31, taking each fused data in the time series as a node in the futures network; S32, for each fused data in the time series, judging whether there is a connection between two fused data through a visualization formula; S33, if there is a connection between the two fused data, establishing an edge between the two fused data; S34, repeating S32~S33 to construct a futures network.

5. The investment risk control method based on futures network according to claim 4 is characterized in that: The visualization formula is as follows: ； Set the number of consecutive trading days T is the time variable, , the fused data set under the time variable is recorded as X , ， ,b,c∈n, for and Any point between Satisfies the visualization formula, where , indicating the fusion data set X Fusion Data and x b A connection exists.

6. The investment risk control method based on futures network according to claim 1 is characterized in that: The centrality indicators include degree value, closeness centrality, harmonic closure centrality, betweenness centrality, PageRank value, eigenvector centrality, i≤n ; The degree value is calculated as follows: ； Is a node i The degree value, Is a node i and nodes j If the edge between nodes i and nodes j If there is an edge between The value of is defined as 1, otherwise it is 0; The closeness centrality is calculated as follows: ； Is a node i The closeness centrality, n is the number of nodes in the futures network, Is a node i To Node j The shortest path length; The harmonic closure centrality is calculated as follows: ； Is a node i The harmonic closure centrality of n is the number of nodes in the futures network, Is a node i To Node j The shortest path length; The betweenness centrality is calculated as follows: ； Is a node i The betweenness centrality of , v, s, t all represent nodes, It is a slave node s To Node t The number of shortest paths, It is a slave node s To Node t The shortest path passes through the node i The number of shortest paths; The PageRank value is calculated as follows: ； Is a node i In the t The PageRank value of the iteration, Is a node i In the t-1 The PageRank value of the iteration, is an indicator function, if the node j To Node i If there is a direct connection, = 1, otherwise = 0, For Node j The out-degree of The eigenvector centrality is calculated as follows: ； Is a node i The eigenvector centrality of is the adjacency matrix of the futures network, λ is the largest eigenvalue of the adjacency matrix, Is a node i The corresponding feature vector.

7. The investment risk control method based on futures network according to claim 6 is characterized in that: In S5, the centrality index of each node is fused based on the DS evidence theory to obtain the importance characterization value of each node, and the nodes are sorted from large to small according to the importance characterization value. The top N nodes in the sorting are recorded as key risk nodes, and the remaining nodes are recorded as non-key risk nodes.

8. The investment risk control method based on futures network according to claim 7 is characterized in that: Based on the DS evidence theory, the normalized values ​​of the centrality index of each node are fused to obtain the importance characterization value of each node, including the following operations: S51, calculate the normalization coefficient 1- K : ； S52, according to the normalization coefficient 1- K Calculate the importance of each node ; 。

Images