Research on Cross-platform Data Interaction and Information Fusion Mechanism of Public Security Intelligence for National Security

Public security intelligence work has always been an important part of public security work. In the early days, intelligence work was dominated by human intelligence activities. With the advancement of public security informatization construction and the deepening of public security organs to implement intelligence-led policing strategy, and vigorously promote the construction of public security intelligence system, public security intelligence work has entered into a period of comprehensive development [1-4]. The development and utilization of various information resources, from which to discover effective intelligence, has become the central task of the current public security intelligence work, the formation of a new public security intelligence work mode - information resources development model [5-7].

This working mode not only conforms to the development trend of public security informatization, promotes the high-end development of public security informatization construction, realizes the close integration of public security intelligence work and public security informatization construction, but also contributes to the development and innovation of public security intelligence work [8-11]. However, due to the public security organs of information resources acquisition, integration and other aspects of the existence of a number of drawbacks and problems, and constraints on the development of this mode of intelligence work. First, too much reliance on the source of information collection, resulting in grass-roots police information collection workload [12-15]. Secondly, information processing and integration face many difficulties. Third, the timeliness of information resources is difficult to be guaranteed. For cross-police, cross-departmental business collaboration, the establishment of public security service platform, based on the public security comprehensive resource base to achieve the public security internal and external data exchange and standardization of service interfaces, to provide cross-police business collaboration, data sharing services is an important direction of development of public security intelligence in the information age [16-19].

In this paper, a cross-platform data interaction method based on a relay chain is proposed, which can meet the requirements of massive data flow and ensure the security and global consistency of cross-platform data transmission.Furthermore, this paper proposes a cross-chain security constraint algorithm.Next, the problem of parallel processing of intelligence distribution based on an information grid is studied, a hierarchical scheduling method is proposed, and the parallel processing flow of information fusion based on an information grid is designed. The article conducts performance tests on the data interaction and information fusion methods proposed in this paper, tests the data security after applying the methods proposed in this paper, and further evaluates the application effect of the cross-platform data interaction and information fusion mechanism proposed in this paper for national security public security intelligence.

2

Method

2.1

Public Security Intelligence for National Security

With regard to the national security concept, the current literature mainly interprets the national security concept in terms of its background, historical evolution and theoretical connotations in the light of the situation of counter-terrorism and terrorism prevention, the soundness of the legal system, the reform of intelligence work and other specific areas. To summarize, the national security concept talks a lot about the concern for non-traditional security, which fully demonstrates that the new national security concept is inclusive and absorbing of the current security situation and security challenges, and that it is a foundational security theory in line with the development of the times, which determines China’s security governance model. The exploration of the relationship between intelligence work and national security is a gradual process: intelligence work is an important means of maintaining national security, and with the changes in the social situation, the legal status of intelligence work should be improved and intelligence work should be integrated into the national security system. Then proposed to do a good job of maintaining non-traditional security, the key is to establish a sound intelligence early warning mechanism. And then studied the relationship between intelligence work and non-traditional security, and proposed that intelligence work can be used to maintain non-traditional security deficiencies. Nowadays, under the guidance of the new national security concept, more and more researches are centered on the intrinsic requirements of the new national security concept on intelligence work and how to make changes in intelligence work.

2.2

Implementation of cross-platform protocols based on relay chains

This paper aims to design a cross-platform data interaction and information fusion mechanism for public security intelligence. Before the application chain executes the cross-platform protocol, it needs to register relevant information on the relay chain. This is the premise and prerequisite for initiating cross-platform public security intelligence. The registered information mainly consists of two parts: the information about the application chain itself and the information about the smart contract deployed on the application chain, which is related to cross-platform. These messages are also transmitted to and recorded by the relay chain in json format [20]. The specification for the application chain registration message is shown in Table 1.

Table 1.

Application chain registration message specification

Parameter	Remark
App ChainID	Application chain id
Type	Application chain type
Certificate	The letter of the endorsement node is the letter of the letter
CBP	The CBP contract address deployed on the application chain
Extra	Other information

In the above parameters, AppChainID is the unique identifier and primary key of the application chain on the relay chain, Type parameter identifies the type of the application chain, Certificate indicates the information of the endorsing node, CBP identifies the address of the CBP contract deployed on the application chain. The registration message also retains Extra for further expansion.

The message specification of the smart contract is shown in Table 2. AppChainID is the ID of the application chain where the smart contract is located, and ChaincodelD is the ID of the smart contract, which is the primary key of the smart contract.

Table 2.

Intelligent contract registration message specification

Parameter	Remark
Appchainid	Application Chain Id
Chaincodeld	Service Id
Type	Service Type
Policy	<Privacy Type>:<Chainld+Chaincodeld>
Extra	Other Information

The cross-chain transaction information structure is shown in Figure 1. Firstly, the initiating chain initiates the public security intelligence cross-platform tx1 and submits it to the relay chain. The relay chain verifies whether the transaction is legal, i.e., whether it has been agreed on the initiating chain and recorded into the world state.

2.3

Security access control for public security intelligence data

2.3.1

Encryption communication protocol design

1)

Description of protocol-related parameters

The protocol is divided into five phases: communication request phase, relay chain response phase, key generation phase, session key generation phase and data transmission phase. Among them, the communication request phase is the data transmission request initiated by the application chain to the relay chain. The relay chain response phase is when the relay link receives the request and prepares the relevant parameters for generation [21].

2)

Communication request phase.

The application chain sends a link establishment request to the relay chain, which contains the ID information Appchain_ID registered on the relay chain in the early stage of the application chain, the certificate CertificateAC returned to the application chain by the relay chain, a random number Random_AC generated by the client of the application chain, and the unique identifier Session ID of this session. The phase of communication requests is shown in Figure 2.

3)

The relay chain response phase is shown in Figure 3.

After the relay links to the application chain request, it first checks the legitimacy of the application chain certificate. After confirming its legitimacy, a random number is generated random_CC. An elliptic curve is selected and the hexadecimal group associated with it is determined (q,G,a,b,n,h). A random number generator is used to generate the integer d[1,n–2] and the point P is calculated according to the following formula: (1) $P = (x_{p}, y_{p}) = [d] G$

The public-private key pair 〈sm2PubKey,sm2PriKey〉 is derived, where the public key sm2PubKey = P and the private key sm2PriKey = d. After generating the public-private key, SessionID is written to the relay chain through the smart contract SC_upload. The relay chain sends the following three messages to the application chain: ① Request Response TxResponse:〈Random_CC, SessionID〉. Where Random_CC is a random number generated by the relay chain. SessionID is the ID of this session. ② Key exchange information CCkKeyEachange:〈sm2PubKeyLength,sm2PubKey〉. which contains the public key length information and public key information on the relay chain. ③ End-of-message-transmission flag TxResponse Done. When the sending is finished, it is CC Response finished.

4)

The key generation phase is shown in Fig. 4.

The application chain generates a pre-master key Pre_master, which is encrypted with the public key of the relay chain into a ciphertext C_{Pre_master} = SM2(Pre_master,sm2PubKey). The encryption method is as follows, where M is the form of Pre_master after it has been converted into a bit string, and klen is the bit length of M: Generate a random number k∈[1,n–1], and compute the elliptic curve point C₁ according to the following formula, and convert its type into a bit string: (2) $C_{1} = [k] G = (x_{1}, y_{1})$

Calculate the elliptic curve point S: (3) $S = [h] P$

Calculate the following elliptic curve points and convert the data types of x₂, y₂ to bit strings: (4) $[k] P = (x_{2}, y_{2})$

Calculate t, C₂ and C₃ according to the following equations: (5) $t = K D F (x_{2} ‖ y_{2}, k l e n)$ (6) $C_{2} = M \oplus t$ (7) $C_{3} = H a s h (x_{2} ‖ M ‖ y_{2})$

The final output is the ciphertext: (8) $C_{\Pr e_m a s t e r} = C_{1} ‖ C_{2} ‖ C_{3}$

When encryption is complete, the message is encapsulated as an ACKey Exchange and sent to the relay chain. At the same time, a Change Cipher Spec message is sent to inform the relay chain that subsequent communications are carried out using session key encryption.

5)

The session key generation phase is shown in Fig. 5.

The application chain and the relay chain use the random numbers generated in the previous stage and the pre-master key to generate the session key, respectively. The relay chain needs to decrypt the received ciphertext C according to the private key, and the decryption process is as follows:

Take out bit string C₁ from ciphertext C_{Pre_master}, convert it into points on the elliptic curve, verify whether it satisfies the elliptic curve equation, and report an error to exit if it does not. Calculate the elliptic curve point S according to the following formula: (9) $S = [h] C_{1}$

Calculate point [d]C₁ by the following formula and convert it to a bit string: (10) $[d] C_{1} = (x_{2}, y_{2})$

Calculate t according to the following formula to confirm that it is not 0: (11) $t = K D F (x_{2} ‖ y_{2}, k l e n)$

Take bit string C₂ from C and compute the plaintext according to the following formula: (12) $M' = C_{2} \oplus t$

Verify the integrity of the plaintext by checking whether u, which is derived from the following formula, is equal to the bit string C3 in ciphertext C: (13) $u = H a s h (x_{2} ‖ M ‖ y_{2})$

6)

Data transfer phase

The application chain encrypts the file using the session key and then transmits the file to the relay chain, which transmits it to the IPFS system and records the encrypted file digest on the relay chain via SC_upload.

2.3.2

Cross-platform security constraint algorithms

This mechanism allows each application chain to set its own visibility policy, i.e., which of its own data is allowed to be queried, forwarded, and saved by Public Security Intelligence across platforms on the relay chain. At the same time, there is a clear order of priority for visibility between different blockchain application systems [22]. The source owner of a file refers to the blockchain system from which the file originally originated, which should have the highest privilege over the blockchain’s sharing policy. The source owner of the file should not lose control of the file when it is forwarded to other blockchain systems. The data transfer phase is shown in Figure 6.

Privacy Leakage due to Visibility Policy Conflict As shown in Figure 7, blockchain system A needs to share files across the chain to the relay chain in the actual business process, and at the same time, the privacy policy set includes blockchain systems B, C, and D, but does not want the files to be forwarded, viewed, or saved by other blockchain systems. However, the sharing policy set by blockchain system D includes other blockchain systems. After being saved by blockchain system D, blockchain system E and other blockchain systems that should not have access to the file have accessed the file’s information, resulting in a data leak from blockchain system A. The propagation path of the upper left arrow is a legitimate path that complies with the privacy policy of blockchain system A, while the lower propagation path poses a security risk to the privacy breach of blockchain system A.

2.4

Cross-platform information fusion for public security intelligence

2.4.1

Movement control model

The public security intelligence processing task scheduling based on the information grid is shown in Figure 8. A~D is the task scheduler, which can be doubled by the main processing node of each processing node of the platform. 1 to 9 are processing nodes, which can be set to become the master node (management and control node), the first-level fusion slave node, and the second-level fusion slave node, as needed, when the fusion parallel processing is completed. The task scheduler is divided into multiple layers. The scheduler of a single platform implements scheduling for the resources of this platform, and when the resources of this platform are insufficient to complete the information processing work, the scheduler will make a request to the scheduler of the previous level, and the higher-level scheduler will then call other resources in the region to participate in the processing, and the layered scheduling is easy to realize the expansion, fault-tolerance, and collaborative allocation, but it does not support the resource geographic autonomy and multiple scheduling strategies.

2.4.2

Movement control methods

The complexity of the scheduling mechanism mainly includes: the time complexity of the scheduling algorithm and the communication overhead in parallel computation, effective scheduling of the tasks executed on the grid and the communication between the tasks is the key for the whole scheduling mechanism to meet the performance requirements, in this paper, we mainly take into account the above factors, and design a multi-platform collaborative task scheduling algorithm based on the improved Back-filling algorithm.

Step 1 A platform I generates a new information processing task, and the task scheduler of this public security intelligence platform completes the task slice processing to form ξ task slices. The scheduler adds task slices to the task queue, re-queues the task queue according to the processing priority, and the task slices of the same information processing task are at the same priority.

Step 2 The task scheduler G_SI of this public security intelligence platform extracts an information processing task from the head of a task queue, and looks up a collection R_s = {r_s1,r_s2,⋯r_sn} of the set of resources of this platform in a resource database that satisfies the resource requirement.

Step 3 When ξ ≤ i, calculate whether the information processing delay of the single platform meets the information processing requirements, and the information processing delay of the single platform T = T_P+T_s+T_c, where: T_P is the task processing delay. T_s is the task scheduling time delay. T_c is the communication delay.

a. Task processing delay (14) $T_{P} = \max t_{i j}; t_{i j} = w (a_{i}) / S_{j} + B_{j}$ Where: t_ij denotes the execution time of task slice a_i on processor P_j. w(a_i) denotes the workload size of the task slice. S_j is the execution speed of processor P_j. B_j is the time required to start a process execution on processor P_j. b. Task scheduling delay T_l = w(a)/S_I+B_l, where: T_I denotes the scheduling delay of task a at the I th scheduler. w(a) denotes the workload for scheduling task a. S_I is the execution speed of the task scheduler I. B_I is the time required to start the execution of a process on scheduler I. c. Communication delay T_c = maxtc_i. τ_{C_i} = (d_i/M+τ)×H_j+O_j, where: tc_i denotes the communication overhead between the processing node P_j executing task slicing a_i on the node and the task scheduler. M denotes the message transfer rate between the task scheduler and each processing node, and is assumed t_{c_i} = M. d_i to denote the amount of data that the task slice a_i interacts with the task scheduler in messages during execution. τ denotes the message passing start time of each processing node and is also assumed to be equal. H_j denotes the number of lines between the processing node P_j and the task scheduler. O_j denotes the contention overhead of the communication lines between the processing node P_j and the task scheduler, forming a mapping between the processing node that satisfies the processing delay and the task slice, and sending the task slice to that processing node for processing, otherwise turning to step 5.

Step 4 If ξ > i, select i tasks to be processed in this platform, and turn to step 3 to see if the processing delay requirement is satisfied.

Step 5 If the processing delay does not meet the requirement, and the public security intelligence platform still has free resources to meet the resource requirement, the processing nodes in the set R_s = {r_s1,r₂,⋯,r_sn} of the resource set of the public security intelligence platform that have already formed a mapping and the processing nodes that cannot meet the processing delay are eliminated to form a new set R_s = {r_r1,r_s2,⋯,r_sm} of the resource set of this platform, and turn to step 3 to calculate whether the other slices in turn meet the processing delay.

Step 6 shifts to step 7 for cross-platform scheduling for task slices that cannot be processed because there are no free resources on this platform to meet the resource requirements.

Step 7 makes a resource request to the higher-level task scheduler (global scheduler). The global scheduler looks up the set of resource sets from each platform in the resource database that satisfy the resource demand R_s.

Step 8 calculates whether the information processing delay of the public security intelligence cross-platform meets the information processing requirements, and the information processing delay of the public security intelligence cross-platform is T_K = T_P+T_S+T_C. Assuming that the scheduling delay is the same at each level, N is the scheduling level, and the total scheduling delay is: T = T_I×(N+1), where T_I is given by b in Step 3.

The communication delay T_C = tc_IJ(N+1)+maxtc_{I_i}, where: tc_IJ = (d_IJ/M₀+τ₀)×H_IJ+O_IJ. tc_{t_i} = (d_i/M_I+τ)×H_j+O_j, tc_IJ is the communication overhead between the scheduler I and the scheduler J, which is assumed to be the same between every two schedulers in the same processing task in this paper. tc_{t_i} denotes the communication overhead between the processing node P_j on the platform I and the task scheduler of this platform while executing the task slice a_i. M₀ denotes a message transfer rate between each of the two task schedulers to each other, and it is assumed that the message transfer rate between each of the two task schedulers is equal. d_IJ denotes the amount of data for message interaction between the scheduler I and the scheduler J during the scheduling of the task a. τ₀ denotes the message passing start time for each scheduler, which is also assumed to be equal. H_IJ denotes the number of lines between scheduler I and scheduler J. O_IJ denotes the contention overhead of the communication lines between scheduler I and scheduler J. M_I denotes the message transfer rate between the task scheduler on platform I and the processing nodes on this platform. d_i denotes the amount of data that the task slice a_i on platform I interacts with the task scheduler in messages during execution. τ denotes the message passing start time for each processing node on a single platform, which is also assumed to be equal. H_j denotes the number of lines between the processing node P_j and the task scheduler on the platform I.

Step 9 maps the processing node that satisfies the processing delay to the task slice, and sends the task slice to the processing node for processing, otherwise turn to step 10.

Step 10 If the processing delay does not meet the requirements, remove the processing nodes that have formed a mapping in the set R_s of public security intelligence cross-platform resource sets and the processing nodes that cannot meet the processing delay, and form a new set R_s = {r_s1,r_s2,⋯,r_sm} of public security intelligence cross-platform resource sets. Turn to step 8 to process the other task slices, and when all of the task slices have formed a mapping with the corresponding processing resources, send the task slices to the processing node for processing, and the scheduling round ends. This round of scheduling ends.

2.4.3

Rasterized data fusion processing flow

The overall process of data fusion parallel processing for rasterization is divided into the following steps: a. Judge the amount of fusion processing jobs and form the set of node resources participating in the fusion processing. b. Derive the fusion processing application on the raster node according to the specified configuration. c. Register the fusion processing tasks after the application process is started. d. Perform the initialization of the user application. e. Gather the node machine names of all the other fusion process nodes and the related resource information. f. Judge the node type of this process according to the discriminating rules, in terms of processing node allocation, there is only one master node, several primary fusion slave nodes, and one secondary fusion slave node.g. According to the node type, sub-tasks of this fusion process are derived. The sub-tasks on the master node are: intelligence network reception task, information source target management task, intelligence division task, parallel computing scheduling, and work status display task, and the sub-tasks on the fusion slave node are: intelligence network reception task, information source target management task, and data fusion processing task. h. Enter into a message event processing loop, and terminate the sub-tasks that this fusion process has derived subtasks. The subtasks release occupied system resources and other cleanup work before termination.i. End the operation of the data fusion processor.

3

Results and discussion

3.1

Cross-platform data interaction and information fusion experimental test analysis

3.1.1

Data interaction experiments

In order to test the application performance of the design platform in realizing the cross-platform information interaction and sharing of public security intelligence, simulation tests are carried out, the time interval of the feature collection of the cross-platform information of public security intelligence is 100 s, the length of the big data distribution is 2000, the size of the feature training set of the cross-platform information of public security intelligence is 500, and the width of the code element of the intelligent scheduling of the cross-platform information of public security intelligence is 0.12 ms. According to the above-mentioned simulation parameter settings, simulation experiments are carried out. The public security intelligence data interaction results are shown in Figure 9. From the figure, it can be seen that the method presented in this paper for cross-platform data interaction in public security intelligence is more effective.

The comparison of real-time and recall of cross-platform data interaction for public security intelligence is shown in Figure 10. The real-time and information recall capabilities of various methods for public security intelligence information interaction and sharing are tested, and comparison results are obtained. From the figure, the average information recall of the data interaction method proposed in this paper is 9.11%, while the recall of the other two reference methods (OpenMP and MPI) is 5.74% and 3.12%, respectively. Therefore, the cross-platform data interaction for public security intelligence proposed in this paper can accurately realize the adaptive scheduling of cross-platform data for public security intelligence, with higher information recall and better checking accuracy.

3.1.2

Comparison of information fusion accuracy rates

The accuracy of information fusion under other variable loads is shown in Fig. 11. The figure shows the statistical graph of training the model with other loads and testing it with another load.The analysis of the experimental results shows that the average accuracy of the fusion method proposed in this paper is 0.9842.The fusion method still has a high accuracy rate under a large number of experiments with variable loads, which verifies the validity and feasibility of the fusion method that is also applicable in the case of variable loads.

3.2

Comparison of data security

The security comparison is shown in Table 3. The table shows the security comparison between the data interaction and information fusion mechanism proposed in this paper and BH, RLL, and DCD, where “Y” represents that the mechanism has the corresponding attributes and “N” represents that the mechanism does not have the corresponding attributes. As can be seen from the table, the mechanism proposed in this paper has verified the validity of the ciphertext in the equivalence test phase, which is not available in other mechanisms.

Table 3.

Security comparison

Scheme	The Test Phase Examines The Validity Of The Information	Safety	Difficulty
Bh	N	Ow-Id-Cca	Bdh
Rll	N	Ow-Id-Cca&Ow-Cca	Dlp
Dcd	N	Ow-Id-Cca&Ind-Id-Cca	Cdh&Ddh
Ours	Y	Ow-Id-Cca&Ind-Id-Cca	Cdh&Ddh

In this section, we measure privacy security based on a sampling approach, utilizing the maximum and average values of attack success probabilities. To maintain the fairness of random sampling, we take the average of the maximum and average values over 10 iterations. The privacy risk analysis is shown in Table 4. The data from the table shows that the proposed data interaction and information fusion mechanism can ensure the privacy security of public security intelligence data.At the same time, we also notice that the privacy risk of the MovieLens dataset is higher than that of the Amazon dataset. This is because the MovieLens dataset is denser and its corresponding users and items have more overlap, while Amazon is a sparse dataset involving multiple domains with no overlap between items in different regions. Therefore, in a sense, the cross-platform data interaction and information fusion mechanism proposed in this paper can provide better data security.

Table 4.

Privacy risk analysis

Node Num	Movielen Dataset		Amazon Dataset
Node Num	Max	Mean	Max	Mean
100	0.152	0.056	0.04	0.016
1000	0.162	0.053	0.009	0.0065
10000	0.114	0.038	0.0059	0.00455

3.3

Evaluation of public security intelligence data management capacity

3.3.1

Construction of an evaluation system for public security intelligence data management capabilities

Intelligence evaluation is a comprehensive assessment of the authenticity and usability of the information produced, and plays a very important role in intelligence work. In terms of the function of emergency intelligence, the evaluation of intelligence should focus on the four functional elements of the intelligence system, that is, the evaluation of the “organizational structure” under the operating mechanism, “intelligence personnel” using “intelligence platform” to transform “intelligence sources” into “decision-making intelligence”. Under the operating mechanism of “organizational structure”, “intelligence personnel” use “intelligence platform” to transform “intelligence source” into “effect” of decision-making intelligence. This results in an intelligence evaluation system consisting of four evaluation factors: intelligence personnel, organizational structure, intelligence platform (technology), and the effect of final intelligence. The effect evaluation system after using the cross-platform data interaction and information fusion mechanism proposed in this paper is shown in Table 5.

Table 5.

Effectiveness evaluation system

Target Layer	Factor Layer	Index Layer
Public Security Intelligence Evaluation (W)	Intelligence Effect(W1)	Veracity(W11)
		Comprehensiveness(W12)
		Timeliness(W13)
		Novelty(W14)
		Applicability(W15)
	Organization(W2)	Reasonableness(W21)
	Organization(W2)	Coordination(W22)
	Intelligence Officer(W3)	Intelligence Sensitivity(W31)
		Intelligence Level(W32)
		Intelligence Ability(W33)
	Information Platform(W4)	Technical Advance(W41)
		Software Quality(W42)
		Economy(W43)
		Ease Of Use(W44)

3.3.2

Evaluation indicator weights

This section utilizes the Delphi method, which is commonly used in indicator evaluation, to ask experts to review the indicator system and determine the weight coefficients of the indicators. The weights of the evaluation indicators are shown in table 6.

Table 6.

Index weight

Factor Layer	Weighting	Index Layer	Weighting
(W1)	0.36	(W11)	0.26
		(W12)	0.15
		(W13)	0.22
		(W14)	0.2
		(W15)	0.17
(W2)	0.15	(W21)	0.47
(W2)	0.15	(W22)	0.53
(W3)	0.23	(W31)	0.39
		(W32)	0.32
		(W33)	0.29
(W4)	0.26	(W41)	0.22
		(W42)	0.18
		(W43)	0.36
		(W44)	0.24

Tier 1 indicators: $W = (w_{1}, w_{2}, w_{3}, w_{4}) = (0.36, 0.15, 0.23, 0.26)$

Tier 2 indicators: $W_{1} = (w_{1}_{1}, w_{1}_{2}, w_{1}_{3}, w_{1}_{4}, w_{1}_{5}) = (0.26, 0.15, 0.22, 0.2, 0.17)$ $W_{2} = (w_{2}_{1}, w_{2}_{2}) = (0.47, 0.53)$ $W_{3} = (w_{3}_{1}, w_{3}_{2}, w_{3}_{3}) = (0.39, 0.32, 0.29)$ $W_{4} = (w_{4}_{1}, w_{4}_{2}, w_{4}_{3}, w_{44}) = (0.22, 0.18, 0.36, 0.24)$

3.3.3

Establishment of a rubric

In view of the characteristics of the effect evaluation after using the cross-platform data interaction and information fusion mechanism of public security intelligence and the selection of evaluation indexes, the set of comment sets is set as C={c1(very good), c2(better), c3(general), c4(worse), c5(very bad)}, which is expressed mathematically and quantitatively as C={5,4,3,2,1}, and the content of the set of comment sets is explained as shown in Table 7.

Table 7.

The comments are illustrated

Commentary set	Explanation
c1 (good)	The criteria and requirements of the index are fully met
c2 (better)	In conformity with most indicators, only individuals can’t meet the requirements
c3 (general)	Basic compliance criteria, less projects that can’t meet the requirements
c4 (worse)	The basic discrepancies are not able to meet the requirements of the project
c5 (Very bad)	The vast majority of the indicators are inconsistent with the criteria, and only individuals can meet the requirements

Considering the ease of operation and the representativeness of the experts, we distributed questionnaires to a total of 10 experts in the field of intelligence and security, who have both some practical work experience and deep theoretical training, and whose opinions are quite representative, and asked them to score the various situations of public security intelligence data management capabilities, and then executed specific evaluations using the fuzzy comprehensive evaluation method. Since the experts’ comments on the public security intelligence data management capability were not very uniform, the weighted average method was used rather than the maximum affiliation method for calculation. The specific evaluation grades are shown in Table 8.

Table 8.

Specific evaluation grade

Index layer	c1	c2	c3	c4	c5
(W11)	0.5	0.3	0.1	0.1	0
(W12)	0.4	0.2	0.2	0.1	0.1
(W13)	0.4	0.3	0.2	0.1	0
(W14)	0.4	0.3	0.3	0	0
(W15)	0.3	0.4	0.1	0.1	0.1
(W21)	0.5	0.1	0.2	0.2	0
(W22)	0.3	0.2	0.2	0.1	0.2
(W31)	0.6	0.3	0.1	0	0
(W32)	0.4	0.5	0.1	0	0
(W33)	0.5	0.3	0.1	0.1	0
(W41)	0.4	0.3	0.2	0.1	0
(W42)	0.6	0.1	0.2	0.1	0
(W43)	0.5	0.3	0.2	0	0
(W44)	0.3	0.2	0.4	0.1	0

The evaluation matrix is: $R_{1} = [\begin{matrix} 0.5 & 0.3 & 0.1 & 0.1 & 0 \\ 0.4 & 0.2 & 0.2 & 0.1 & 0.1 \\ 0.4 & 0.3 & 0.2 & 0.1 & 0 \\ 0.4 & 0.3 & 0.3 & 0 & 0 \\ 0.3 & 0.4 & 0.1 & 0.1 & 0.1 \end{matrix}]$ $R_{2} = [\begin{matrix} 0.5 & 0.1 & 0.2 & 0.2 & 0 \\ 0.3 & 0.2 & 0.2 & 0.1 & 0.2 \end{matrix}]$ $R_{3} = [\begin{matrix} 0.6 & 0.3 & 0.1 & 0 & 0 \\ 0.4 & 0.5 & 0.1 & 0 & 0 \\ 0.5 & 0.3 & 0.1 & 0.1 & 0 \end{matrix}]$ $R_{4} = [\begin{matrix} 0.4 & 0.3 & 0.2 & 0.1 & 0 \\ 0.6 & 0.1 & 0.2 & 0.1 & 0 \\ 0.5 & 0.3 & 0.2 & 0 & 0 \\ 0.3 & 0.2 & 0.4 & 0.1 & 0 \end{matrix}]$

Comprehensive evaluation: $B_{1} = W_{1} * R_{1} = (0.409, 0.302, 0.177, 0.08, 0.032)$ $B_{2} = W_{2} * R_{2} = (0.394, 0.153, 0.2, 0.147, 0.106)$ $B_{3} = W_{3} * R_{3} = (0.507, 0.364, 0.1, 0.029, 0)$ $B_{4} = W_{4} * R_{4} = (0.448, 0.24, 0.248, 0.064, 0)$

All results are used as the evaluation transformation matrix for the first level, so there is: $R = [\begin{matrix} 0.409 & 0.302 & 0.177 & 0.08 & 0.032 \\ 0.394 & 0.153 & 0.2 & 0.147 & 0.106 \\ 0.507 & 0.364 & 0.1 & 0.029 & 0 \\ 0.448 & 0.24 & 0.248 & 0.064 & 0 \end{matrix}]$ $B = W * R = (0.439, 0.278, 0.181, 0.074, 0.027)$

If each level in the evaluation set is quantized as c = (5, 4, 3, 2, 1), the comprehensive evaluation results of the management capability of public security intelligence data after the application of the cross-platform data interaction and information fusion mechanism of public security intelligence proposed in this paper are: $B * c = (0.439, 0.278, 0.181, 0.074, 0.027) * (\begin{matrix} 5 \\ 4 \\ 3 \\ 2 \\ 1 \end{matrix}) = 4.028$

It can be seen that the public security intelligence cross-platform data interaction and information fusion mechanism proposed in this paper for national security is utilized to evaluate the management capability of public security intelligence data as good.

4

Conclusion

It is imperative to utilize cross-platform data interaction and information fusion mechanisms for the maintenance of national security-oriented public security intelligence. Based on this, the article analyzes data interaction technology and information fusion technology, and draws the following conclusions:

In the experimental test analysis, the average information checking rate of the data interaction method proposed in this paper is 9.11%, which is 3.37% higher than the average information checking rate of the other two reference methods, while the checking rates of the other two reference methods are 5.74% and 5.99% respectively. In the comparative analysis of data security, the mechanism proposed in this paper verifies the validity of the ciphertext during the equivalence test phase. By applying the fuzzy comprehensive evaluation method to evaluate the utilization effect of the cross-platform data interaction and information fusion mechanism for public security intelligence proposed in this paper, the evaluation result obtained is good.

Acknowledgements

National Key Research and Development Program (2023YFC3321604);

Special Fund Project for Basic Scientific Research Business Expenses of Central Universities (2022JKF02002);

Double First-Class Innovation Research Special Project of the People’s Public Security University of China (2023SYL20).

Langue:: Anglais

Périodicité:: 1 fois par an
Sujets de la revue:: Sciences de la vie, Sciences de la vie, autres, Mathématiques, Mathématiques appliquées, Mathématiques générales, Physique, Physique, autres

RSS Feed de la revue

Research on Cross-platform Data Interaction and Information Fusion Mechanism of Public Security Intelligence for National Security

Haohang Ye

Wengang Feng

Publié en ligne: 17 mars 2025

Reçu: 22 oct. 2024

Accepté: 02 févr. 2025

DOI: https://doi.org/10.2478/amns-2025-0309

Mots clésData interaction, Information fusion, Public security intelligence, Information grid

© 2025 Haohang Ye et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Mots clés
Data interaction, Information fusion, Public security intelligence, Information grid