A method for constructing a smart home quality risk assessment model based on big data

Abstract

The present invention discloses a method for constructing a smart home quality risk assessment model based on big data, comprising the steps of acquiring parameter data and communication data of the smart home, obtaining a first assessment score based on the parameter data, performing data security analysis based on the communication data, obtaining vulnerability availability and compatibility, obtaining an attack prediction path based on resource game, performing risk quantification on the attack prediction path to obtain vulnerability priority, calculating data security according to vulnerability priority and vulnerability availability, calculating a second assessment score according to data security and compatibility, further constructing a quadrant assessment system, constructing a quality risk assessment model based on a neural network, and using the quadrant assessment system as an assessment benchmark for the quality risk assessment model. The method can not only improve the comprehensiveness and accuracy of smart home quality risk assessment, but also has good interpretability, and can be directly applied to the smart home quality risk assessment system.

Description

Technical Field

[0001] The present invention relates to the field of quality risk assessment, and in particular to a method for constructing a smart home quality risk assessment model based on big data. Background Art

[0002] With the rapid development of the Internet of Things, artificial intelligence and big data, the intelligence and automation levels of smart home systems are constantly improving. Quality risk assessment of smart home devices and systems has become a key factor in ensuring product quality and system safety.

[0003] Traditional smart home quality risk assessment methods mainly rely on simple fault detection and manual experience judgment, but this method has problems such as high cost, low efficiency, and strong subjectivity. In recent years, with the rapid development of technology, more and more smart home assessment methods have begun to use intelligent algorithms or big data for risk assessment to reduce costs and improve the efficiency and accuracy of detection and assessment.

[0004] At present, although there are some quality risk assessment models based on smart devices, the evaluation methods in these assessment models often only focus on evaluating a single dimension or comparing multiple indicators. The evaluation methods are relatively simple and lack a comprehensive assessment of the overall quality. In addition, the evaluation results of existing models are often used for quality control, ignoring the consideration of user experience. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a smart home quality risk assessment model based on big data.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] The first aspect of the present invention provides a method for constructing a smart home quality risk assessment model based on big data, comprising the following steps:

[0008] Step S1, obtaining parameter data of a smart home, and performing data processing rate analysis on the smart home based on the parameter data to obtain a first evaluation score;

[0009] Step S2, acquiring communication data of the smart home, performing data security analysis on the communication data to obtain vulnerability exploitability, and performing interaction simulation on the smart home based on the communication data to obtain smart home compatibility; the communication data includes smart home system transmission data and interaction logs between users and the system;

[0010] Step S3, obtaining an attack prediction path, quantifying the risk of the attack prediction path, obtaining a vulnerability priority based on the quantification result, calculating data security according to the vulnerability priority and the vulnerability exploitability, and calculating a second evaluation score according to the data security and the compatibility;

[0011] Step S4, divide the first assessment score and the second assessment score into quadrants, construct a quadrant assessment system, construct a quality risk assessment model based on a neural network, and use the quadrant assessment system as an assessment benchmark for the quality risk assessment model.

[0012] Furthermore, the method for analyzing the data processing rate of a smart home based on the parameter data includes:

[0013] Calculate the first evaluation score using the formula:

[0014]

[0015] Among them, f c is the data storage rate of the smart home processor, f s is the data storage rate of the smart home memory, ψ is the memory size of the memory, H la is the delay time of data transmission process, H ji is the jitter rate of the data transmission process, dB is the number of concurrent connection channels, and G is the data compression rate.

[0016] Furthermore, the method for performing data security analysis on the communication data includes:

[0017] Construct a data transmission channel to send avalanche data to the smart home system to simulate the attack scenario;

[0018] Build a vulnerability search strategy, obtain the vulnerabilities in the binary file, and calculate the exploitability of each vulnerability. The calculation formula is:

[0019]

[0020] Among them, α represents the frequency of vulnerability discovery, β is the channel capacity of avalanche data, and R n is the bandwidth-delay product of the data transmission channel, W represents the number of function calls in the vulnerability, B is the signal bandwidth of the binary file, and r i is the signal-to-noise ratio between sending and receiving operations of the binary file, and t is the update period of the binary file.

[0021] Furthermore, the method for constructing a vulnerability search strategy includes:

[0022] Obtain a binary file of a data transmission channel in a smart home system, and convert the binary code in the binary file into a disassembled code;

[0023] Obtain the calling function in the disassembled code, wherein the calling function includes: a buffer function and a heap function, and locate the common form and the expanded form in the calling function, wherein the common form is a function called by a call instruction, and the expanded form is a function called by a function pointer, a jump table, and an indirect addressing;

[0024] Observe whether the stack overflows in the normal form and the expanded form of the calling function. If overflow occurs, check whether the overflow length covers the return address. If the stack overflow covers the return address, it is determined that the binary file has a vulnerability. The specific process expression is:

[0025] StackPointer+StackFramSize-basePointer>maxStackSize

[0026] OverflowLength≥addressOfReturn-stackPointer

[0027] Among them, StackPointer is the value of the current stack pointer, StackFramSize is the current stack frame size, basePointer is the starting address of the current stack frame, maxStackSize is the maximum allowed stack usage, OverflowLength is the overflow data length, and addressOfReturn is the memory address of the return address.

[0028] Furthermore, the method for interactively simulating a smart home based on the communication data includes:

[0029] When the smart home system meets the data isolation condition, the compatibility analysis of the smart home is carried out, including the compatibility of human-computer interaction and the compatibility between devices. The overall compatibility of the smart home is calculated based on the communication data and the response accuracy of human-computer interaction. The calculation formula is:

[0030]

[0031] Among them, L N It is expressed as the effective data volume, R is the sending end rate, T is the channel sending period, Q is the channel sending period, os is the channel bandwidth packet loss rate, n e is the number of bit errors transmitted, n is the total number of transmitted codes, S is the total number of channel links, K is the number of links occupied by transmission, L f is the data frame length, H is the channel bandwidth, It is expressed as the response accuracy of smart home human-computer interaction.

[0032] Furthermore, the method of obtaining the attack prediction path, quantifying the risk of the attack prediction path, and obtaining the vulnerability priority based on the quantification result includes:

[0033] The game of resources in the channel based on game theory is used as a simulation process of attacking the attacker's vulnerability and repairing the vulnerability of the smart home. The optimal attack-repair game strategy of the vulnerability is obtained based on the Hasani transformation.

[0034] The complexity of the optimal attack-repair game strategy for calculating the vulnerability is expressed as:

[0035]

[0036] Among them, {n a , <r k ,u i >,∑{[var]}} represents the information carried by the attack, → represents the game access to simulate the attack, {n a , <r z ,u j >,θ _score ,∑{[var:val}} represents the resources of the attack game,θ _score Represents channel r z Vulnerability j Current resource benefits;

[0037] The vulnerability state transition probability is calculated based on the complexity of the optimal attack-repair game strategy of the vulnerability. The calculation formula is:

[0038]

[0039] Where a1, b1 and a2, b2 represent the upper and lower limits of x and y respectively, Φ is the complexity of the attack-repair game strategy of the vulnerability, v(k) represents the exploitable rate of the vulnerability, x and y are the attack cost window and repair cost window of the attack-repair game strategy respectively, and μ is the life cycle factor of the vulnerability;

[0040] The attack path is predicted by the hidden Markov model according to the vulnerability state transition probability to obtain the attack prediction path, and the vulnerability priority is obtained according to the order of the nodes on the attack prediction path.

[0041] Furthermore, the method for calculating data security according to the vulnerability priority and the vulnerability exploitability includes:

[0042]

[0043] Where l is the number of vulnerabilities, i is the i-th vulnerability, is the exploitability of the vulnerability, α i is the priority of the vulnerability, K(x i ,x) represents the network congestion rate of the channel where the vulnerability is located, and b is the bias coefficient.

[0044] Furthermore, the method for calculating the second evaluation score according to the data security and the compatibility includes:

[0045] Obtaining a data security and compatibility importance scoring matrix, determining a membership function formed by a gradient distribution, and obtaining the data security and compatibility membership levels based on the membership function and the scoring matrix;

[0046] The probability hesitant fuzzy average operator is used to aggregate each row of the rating matrix. The expression is:

[0047]

[0048] Among them, i=1,2,...,n represents the user, j=1,2 represents the data security index and compatibility index, represents the union operation of all membership levels, n is the number of users participating in the rating, p ij is the score of the i-th user on the j-th indicator, represents the membership level of the i-th user to the j-th indicator, It means calculating the cumulative probability of each membership level. represents the product of all user membership levels, Indicates the consistency of ratings among all users;

[0049] The aggregation results are normalized to obtain the index weights of compatibility and data security. The second evaluation score is calculated based on the index weights of compatibility and data security. The calculation formula is:

[0050] f2=C·W i +E2·W j

[0051] Among them, C is compatibility, E2 is data security, W is i is the compatibility weight, W j is the weight of data security.

[0052] Furthermore, the method of dividing the first evaluation score and the second evaluation score into quadrants to construct a quadrant evaluation system includes:

[0053] Collect historical smart home parameters and communication data, calculate the historical first evaluation score and the historical second evaluation score, sort them from large to small, and select the lowest score of the top 70% of the historical first evaluation score and the top 80% of the historical second evaluation score as the dividing threshold points of the first evaluation score and the second evaluation score, and obtain V1 and V2 respectively;

[0054] Construct a quadrant assessment system, including:

[0055] The data points with the first evaluation score greater than or equal to V1 and the second evaluation score greater than or equal to V2 are regarded as points in the first quadrant area, and the evaluation level is excellent;

[0056] The data points with the first evaluation score greater than or equal to V1 and the second evaluation score less than V2 are regarded as points in the second quadrant area, and the evaluation grade is good;

[0057] The data points with the first evaluation score less than V1 and the second evaluation score greater than or equal to V2 are regarded as points in the third quadrant area, and the evaluation level is general;

[0058] Data points whose first evaluation score is less than V1 and whose second evaluation score is less than V2 are regarded as points in the fourth quadrant area, and their evaluation level is poor.

[0059] Furthermore, the method for constructing a quality risk assessment model based on a neural network includes: defining a neural network architecture, including an input layer, a hidden layer, and an output layer, wherein the input layer is used to obtain parameter data and communication data of the smart home, the hidden layer includes two logic processing layers, wherein the first logic processing layer is used to calculate a first evaluation score and a second evaluation score, the second logic processing layer is a threshold comparison layer, and the first evaluation score and the second evaluation score calculated by the first logic processing layer are based on a quadrant evaluation system to obtain an evaluation grade, the output layer includes two sub-layers, wherein the first sub-layer is used to output the evaluation grade obtained based on the second logic processing layer, and the second sub-layer is a forward propagation layer used to output the probability distribution of data points in each quadrant obtained by processing the first logic layer using a softmax function.

[0060] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0061] (1) The present invention provides a method for constructing a smart home quality risk assessment model based on big data. By acquiring parameter data and communication data of the smart home, a first assessment score and a second assessment score are calculated, and a quadrant assessment system is constructed based on the first assessment score and the second assessment score. This can more directly reflect the assessment level of the smart home. For a large number of smart home assessment projects, the probability distribution output by the output layer of the smart home risk assessment model can intuitively reflect the distribution of the quality risk assessment level. In addition, the first assessment score and the second assessment score can be used to perform comparative assessment separately or in combination, thereby providing a variety of assessment methods.

[0062] (2) The present invention analyzes user compatibility by analyzing the system interaction log in the communication data. The compatibility includes the compatibility between human-computer interaction and between devices. It emphasizes the importance of user experience in the smart home quality evaluation method, so that the evaluation method is more in line with the user's operating habits and needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 The present invention is a flowchart of the steps of a method for constructing a smart home quality risk assessment model based on big data.

[0064] Figure 2 A quadrant diagram of the quadrant evaluation system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all 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.

[0066] Reference Figure 1 As shown, the present invention provides a method for constructing a smart home quality risk assessment model based on big data, comprising:

[0067] Step S1, obtaining parameter data of a smart home, and performing data processing rate analysis on the smart home based on the parameter data to obtain a first evaluation score;

[0068] In the actual evaluation, the smart home parameter data is obtained, including: processor data storage rate: 3000MB / s, memory data storage rate: 800MB / s, memory size: 16GB, data transmission delay time: 10ms, jitter rate of data transmission process: 2ms, number of concurrent connection channels: 20, data compression rate: 500MB / s. The first evaluation score of the smart home calculated based on the parameter data is score1;

[0069] Step S2, acquiring communication data of the smart home, performing data security analysis on the communication data to obtain vulnerability exploitability, and performing interaction simulation on the smart home based on the communication data to obtain smart home compatibility; the communication data includes smart home system transmission data and interaction logs between users and the system;

[0070] In the actual evaluation, avalanche data is constructed with randomly generated payload data, and the data volume is 1G. Within the preset 60s, an avalanche test is performed on the data transmission channel in the smart home system at a rate of 100Mbps. The binary file of the firmware is obtained from the data transmission channel. The firmware version number is 1.2.3 and the file name is router_firmware.bin. The disassembly tool Ghidra is used to convert router_firmware.bin into disassembly code. In the disassembly code, the buffer function and the heap function are identified, including: strcpy, strcat, sprintf, malloc, free functions. The common form of the buffer function and the format of the function are mov rax, [strcpy_function_ptr]call. The expanded form of rax is used to locate the specific parameters of the expanded form and the common form in the buffer function and the heap function, including: the value of the current frame pointer ESP is 0×8E50, the stack frame size of the current function is 0×50, the stack frame starting position EBP of the current function is 0×8F00, the maximum allowed stack usage is 0×1000, the overflow data length is 0×80, and the memory address of the return address is 0×8F40. For each call, check whether there is a risk of stack overflow;

[0071] In the actual evaluation, a total of 4 vulnerabilities were obtained. Further communication data was obtained, including: the frequency of vulnerability discovery per 1,000 communications, channel capacity, bandwidth, bandwidth-delay product, the number of function calls in the vulnerability, signal bandwidth, signal-to-noise ratio, and binary file update cycle. The exploitability of the 4 vulnerabilities was calculated to be

[0072] In the actual evaluation, the interaction logs between users and the system in the smart home system were obtained, including: effective data volume: 100 data packets, sending end rate: 100Mbps, channel sending cycle: 0.01s, channel bandwidth packet loss rate: 2%, transmission bit error number: 20, total number of transmission codes: 1000000, total number of channel links: 10, number of transmission occupied links: 5, data frame length: 1500 bytes, channel bandwidth: 20Mbps, human-computer interaction response accuracy: 94.5%, and the overall compatibility of the smart home was calculated to be C1;

[0073] Step S3, obtaining an attack prediction path, quantifying the risk of the attack prediction path, obtaining a vulnerability priority based on the quantification result, calculating data security according to the vulnerability priority and the vulnerability exploitability, and calculating a second evaluation score according to the data security and the compatibility;

[0074] In the actual evaluation, the optimal strategy of the attack-repair game is obtained based on game theory, and the complexity of the optimal strategy of the attack-repair game of the vulnerability is further calculated. The vulnerability state transition probability is calculated according to the complexity of the optimal strategy of the attack-repair game of the vulnerability. The attack path is predicted by the hidden Markov model based on the state set and the observation set to obtain the Markov chain. The Markov chain is used as the attack prediction path, and the nodes on the attack prediction path are used as vulnerabilities. The priority of the vulnerability is further obtained based on the order of the path nodes. The priorities of the four vulnerabilities are ω1, ω2, ω3, and ω4, respectively. Based on the vulnerability priorities ω1, ω2, ω3, ω4 and the vulnerability exploitability The data security E2 is calculated, and the scoring matrix of the data security E2 and the compatibility C1 is aggregated and normalized to obtain the weight values ​​of the data security E2 and the compatibility C1, and the second evaluation score score2 is further calculated;

[0075] Step S4, divide the first assessment score and the second assessment score into quadrants, construct a quadrant assessment system, construct a quality risk assessment model based on a neural network, and use the quadrant assessment system as an assessment benchmark for the quality risk assessment model.

[0076] Collect historical smart home parameters and communication data, calculate the historical first evaluation score and the historical second evaluation score, sort them from large to small, and select the lowest score in the top 70% of the historical first evaluation score and the top 80% of the historical second evaluation score as the division threshold points of the first evaluation score and the second evaluation score, respectively, and obtain T1 and T2. According to the threshold points, divide the data points into four quadrants and assign corresponding evaluation grades to obtain the quadrant evaluation system:

[0077] Quadrant 1 (Excellent): The first assessment score ≥ T1 and the second assessment score ≥ T2,

[0078] Second quadrant (good level): The first evaluation score ≥ T1 and the second evaluation score < T2,

[0079] Third quadrant (average level): The first evaluation score < T1 and the second evaluation score ≥ T2,

[0080] Fourth quadrant (poor level): The first evaluation score < T1 and the second evaluation score < T2;

[0081] In the actual evaluation, a neural network architecture is defined, including an input layer, a hidden layer, and an output layer. The input layer is used to obtain the parameter data and communication data of the smart home. The hidden layer contains two logical processing layers. The first logical processing layer is used to calculate the first evaluation score and the second evaluation score. The second logical processing layer is a threshold comparison layer. Based on the first evaluation score and the second evaluation score calculated by the first logical processing layer and the quadrant evaluation system, the evaluation level is obtained. The output layer contains two sub-layers. The first sub-layer is to output the evaluation level obtained based on the second logical processing layer. The second sub-layer is a forward propagation layer used to output the probability distribution of each quadrant obtained by processing the first logical layer using the softmax function. Compare the calculated score1 and score2 with the quadrant evaluation system to obtain an evaluation level of good level, and obtain the probability distributions of each quadrant as φ1, φ2, φ3, and φ4 respectively.

[0082] In this embodiment, the method for analyzing the data processing rate of the smart home based on the parameter data includes:

[0083] Calculate the first evaluation score, and the calculation formula is:

[0084]

[0085] where f c is the data storage rate of the smart home processor, f s is the data storage rate of the smart home memory, ψ is the memory size, H la is the delay time during the data transmission process, H ji is the jitter rate during the data transmission process, dB is the number of concurrent connection channels, and G is the data compression rate.

[0086] In this embodiment, the method for analyzing the data security of the communication data includes:

[0087] Construct an avalanche data to send to the data transmission channel of the smart home system to simulate an attack scenario;

[0088] Build a vulnerability search strategy, obtain each vulnerability, and calculate the exploitable rate of each vulnerability. The calculation formula is:

[0089]

[0090] Among them, α represents the frequency of vulnerability discovery, β is the channel capacity of avalanche data, and R n is the bandwidth-delay product of the data transmission channel, W represents the number of function calls in the vulnerability, B is the signal bandwidth of the binary file, and r i is the signal-to-noise ratio between sending and receiving operations of the binary file, and t is the update period of the binary file.

[0091] In this embodiment, the method for constructing a vulnerability search strategy includes:

[0092] Obtain a binary file of a data transmission channel in a smart home system, and convert the binary code in the binary file into a disassembled code;

[0093] Obtain the calling function in the disassembled code, wherein the calling function includes: a buffer function and a heap function, and locate the common form and the expanded form in the calling function, wherein the common form is a function called by a call instruction, and the expanded form is a function called by a function pointer, a jump table, and an indirect addressing;

[0094] Observe whether the stack overflows in the normal form and the expanded form of the calling function. If overflow occurs, check whether the overflow length covers the return address. If the stack overflow covers the return address, it is determined that the binary file has a vulnerability. The specific process expression is:

[0095] StackPointer+StackFramSize-basePointer>maxStackSize

[0096] OverflowLength≥addressOfReturn-stackPointer

[0097] Among them, StackPointer is the value of the current stack pointer, StackFramSize is the current stack frame size, basePointer is the starting address of the current stack frame, maxStackSize is the maximum allowed stack usage, OverflowLength is the overflow data length, and addressOfReturn is the memory address of the return address.

[0098] In this embodiment, the method for interactively simulating a smart home based on the communication data includes:

[0099] Under the condition that the smart home system meets the full data exchange conditions, the compatibility of the smart home is analyzed, including the compatibility of human-computer interaction and the compatibility between devices. The overall compatibility of the smart home is calculated based on the communication data and the response accuracy of human-computer interaction. The calculation formula is:

[0100]

[0101] Among them, L N It is expressed as the effective data volume, R is the sending end rate, T is the channel sending period, Q is the channel sending period, os is the channel bandwidth packet loss rate, n e is the number of bit errors transmitted, n is the total number of transmitted codes, S is the total number of channel links, K is the number of links occupied by transmission, L f is the data frame length, H is the channel bandwidth, It is expressed as the response accuracy of smart home human-computer interaction.

[0102] In this embodiment, the method of obtaining an attack prediction path, quantifying the risk of the attack prediction path, and obtaining a vulnerability priority based on the quantification result includes:

[0103] The game of resources in the channel based on game theory is used as a simulation process of attacking the attacker's vulnerability and repairing the vulnerability of the smart home. The optimal attack-repair game strategy of the vulnerability is obtained based on the Hasani transformation.

[0104] The complexity of the optimal attack-repair game strategy for calculating the vulnerability is expressed as:

[0105]

[0106] Among them, {n a , <r k ,u i >,∑{[var]}} represents the information carried by the attack, → represents the game access to simulate the attack, {n a , <r x ,u j >,θ _score ,∑{[var:val]}} represents the resources of the attack game,θ _score Represents channel r z Vulnerability j Current resource benefits;

[0107] The vulnerability state transition probability is calculated based on the complexity of the optimal attack-repair game strategy of the vulnerability. The calculation formula is:

[0108]

[0109] Where a1, b1 and a2, b2 represent the upper and lower limits of x and y respectively, Φ is the complexity of the attack-repair game strategy of the vulnerability, v(k) represents the exploitable rate of the vulnerability, x and y are the attack cost window and repair cost window of the attack-repair game strategy respectively, and μ is the life cycle factor of the vulnerability;

[0110] The attack path is predicted by the hidden Markov model according to the vulnerability state transition probability to obtain the attack prediction path, and the vulnerability priority is obtained according to the order of the nodes on the attack prediction path.

[0111] In this embodiment, the method for calculating data security according to the vulnerability priority and the vulnerability exploitability includes:

[0112]

[0113] Where l is the number of vulnerabilities, i is the i-th vulnerability, is the exploitability of the vulnerability, α i is the priority of the vulnerability, K(x i ,x) represents the network congestion rate of the channel where the vulnerability is located, and b is the bias coefficient.

[0114] In this embodiment, the method for calculating the second evaluation score according to the data security and the compatibility includes:

[0115] Obtaining a data security and compatibility importance scoring matrix, determining a membership function formed by a gradient distribution, and obtaining the data security and compatibility membership levels based on the membership function and the scoring matrix;

[0116] The probability hesitant fuzzy average operator is used to aggregate each row of the rating matrix. The expression is:

[0117]

[0118] Among them, i=1,2,...,n represents the user, j=1,2 represents the data security index and compatibility index, represents the union operation of all membership levels, n is the number of users participating in the rating, p ij is the score of the i-th user on the j-th indicator, represents the membership level of the i-th user to the j-th indicator, [1- It means calculating the cumulative probability of each membership level. represents the product of all user membership levels, Indicates the consistency of ratings among all users;

[0119] The aggregation results are normalized to obtain the index weights of compatibility and data security. The second evaluation score is calculated based on the index weights of compatibility and data security. The calculation formula is:

[0120] f2=C·W i +E2·W j

[0121] Among them, C is compatibility, E2 is data security, W is i is the compatibility weight, W j is the weight of data security.

[0122] In this embodiment, if Figure 2 As shown, it is a quadrant schematic diagram of the quadrant evaluation system provided in an embodiment of the present application, and the method of dividing the first evaluation score and the second evaluation score into quadrants to construct the quadrant evaluation system includes:

[0123] Collect historical smart home parameters and communication data, calculate the historical first evaluation score and the historical second evaluation score, sort them from large to small, and select the lowest score of the top 70% of the historical first evaluation score and the top 80% of the historical second evaluation score as the dividing threshold points of the first evaluation score and the second evaluation score, and obtain V1 and V2 respectively;

[0124] Construct a quadrant assessment system, including:

[0125] The data points with the first evaluation score greater than or equal to V1 and the second evaluation score greater than or equal to V2 are regarded as points in the first quadrant area, and the evaluation level is excellent;

[0126] The data points with the first evaluation score greater than or equal to V1 and the second evaluation score less than V2 are regarded as points in the second quadrant area, and the evaluation grade is good;

[0127] The data points with the first evaluation score less than V1 and the second evaluation score greater than or equal to V2 are regarded as points in the third quadrant area, and the evaluation level is general;

[0128] Data points whose first evaluation score is less than V1 and whose second evaluation score is less than V2 are regarded as points in the fourth quadrant area, and their evaluation level is poor.

[0129] In this embodiment, the method for constructing a quality risk assessment model based on a neural network includes: defining a neural network architecture, including an input layer, a hidden layer, and an output layer, wherein the input layer is used to obtain parameter data and communication data of the smart home, the hidden layer includes two logic processing layers, wherein the first logic processing layer is used to calculate a first evaluation score and a second evaluation score, the second logic processing layer is a threshold comparison layer, and the first evaluation score and the second evaluation score calculated by the first logic processing layer are based on a quadrant evaluation system to obtain an evaluation grade, the output layer includes two sublayers, wherein the first sublayer is used to output the evaluation grade obtained based on the second logic processing layer, and the second sublayer is a forward propagation layer used to output the probability distribution of data points in each quadrant obtained by processing the first logic layer using a softmax function.

[0130] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for constructing a smart home quality risk assessment model based on big data, characterized in that: The following steps are involved: Step S1, obtaining parameter data of a smart home, and performing data processing rate analysis on the smart home based on the parameter data to obtain a first evaluation score; Step S2, acquiring communication data of the smart home, performing data security analysis on the communication data to obtain vulnerability exploitability, performing interactive simulation on the smart home based on the communication data to obtain smart home compatibility; The communication data includes smart home system transmission data and interaction logs between users and the system; the method for interactive simulation of smart home based on the communication data includes: When the smart home system meets the data isolation condition, the compatibility analysis of the smart home is carried out, including the compatibility of human-computer interaction and the compatibility between devices. The overall compatibility of the smart home is calculated based on the communication data and the response accuracy of human-computer interaction. The calculation formula is: Among them, L N It is expressed as the effective data volume, R is the sending end rate, T is the channel sending period, Q is the channel sending period, os is the channel bandwidth packet loss rate, n e is the number of bit errors transmitted, n is the total number of transmitted codes, S is the total number of channel links, K is the number of links occupied by transmission, L f is the data frame length, H is the channel bandwidth, It is expressed as the response accuracy of smart home human-computer interaction; Step S3, obtaining an attack prediction path, quantifying the risk of the attack prediction path, obtaining a vulnerability priority based on the quantification result, calculating data security according to the vulnerability priority and the vulnerability exploitability, and calculating a second evaluation score according to the data security and the compatibility; Step S4, divide the first assessment score and the second assessment score into quadrants, construct a quadrant assessment system, construct a quality risk assessment model based on a neural network, and use the quadrant assessment system as an assessment benchmark for the quality risk assessment model.

2. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method for analyzing the data processing rate of a smart home based on the parameter data comprises: Calculate the first evaluation score using the formula: Among them, f c is the data storage rate of the smart home processor, f s is the data storage rate of the smart home memory, ψ is the memory size of the memory, H la is the delay time of data transmission process, H ji is the jitter rate of the data transmission process, dB is the number of concurrent connection channels, and G is the data compression rate.

3. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method for performing data security analysis on the communication data comprises: Construct a data transmission channel to send avalanche data to the smart home system to simulate the attack scenario; Build a vulnerability search strategy, obtain the vulnerabilities in the binary file, and calculate the exploitability of each vulnerability. The calculation formula is: Among them, α represents the frequency of vulnerability discovery, β is the channel capacity of avalanche data, and R n is the bandwidth-delay product of the data transmission channel, W represents the number of function calls in the vulnerability, B is the signal bandwidth of the binary file, and r i is the signal-to-noise ratio between sending and receiving operations of the binary file, and t is the update period of the binary file.

4. The method for constructing a smart home quality risk assessment model based on big data according to claim 3 is characterized in that: The method for constructing a vulnerability search strategy includes: Obtain a binary file of a data transmission channel in a smart home system, and convert the binary code in the binary file into a disassembled code; Obtain the calling function in the disassembled code, wherein the calling function includes: a buffer function and a heap function, and locate the common form and the expanded form in the calling function, wherein the common form is a function called by a call instruction, and the expanded form is a function called by a function pointer, a jump table, and an indirect addressing; Observe whether the stack overflows in the normal form and the expanded form of the calling function. If overflow occurs, check whether the overflow length covers the return address. If the stack overflow length covers the return address, it is determined that the binary file has a vulnerability. The specific process expression is: StackPointer+StackFramSize-basePointer>maxStackSize OverflowLength≥addressOfReturn-stackPointer Among them, StackPointer is the value of the current stack pointer, StackFramSize is the current stack frame size, basePointer is the starting address of the current stack frame, maxStackSize is the maximum allowed stack usage, OverflowLength is the overflow data length, and addressOfReturn is the memory address of the return address.

5. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: A method for obtaining an attack prediction path, quantifying the risk of the attack prediction path, and obtaining a vulnerability priority based on the quantification result includes: The game of resources in the channel based on game theory is used as a simulation process of attacking the attacker's vulnerability and repairing the vulnerability of the smart home. The optimal attack-repair game strategy of the vulnerability is obtained based on the Hasani transformation. The complexity of the optimal attack-repair game strategy for calculating the vulnerability is expressed as: Φ={{n a , <r k ,you i >,∑{[var]}}→{n a , <r z ,you j >,θ _score ,∑{[var:val]}}} Among them, {n a , <r k ,u i ), ∑{[var]}} represents the information carried by the attack, → represents the game access to simulate the attack, {n a , <r z ,u j >,θ _score ,∑{[var:val]}} represents the resources of the attack game,θ _score Represents channel r z Vulnerability j Current resource benefits; The vulnerability state transition probability is calculated based on the complexity of the optimal attack-repair game strategy of the vulnerability. The calculation formula is: Where a1, b1 and a2, b2 represent the upper and lower limits of x and y respectively, Φ is the complexity of the attack-repair game strategy of the vulnerability, v(k) represents the exploitable rate of the vulnerability, x and y are the attack cost window and repair cost window of the attack-repair game strategy respectively, and μ is the life cycle factor of the vulnerability; The attack path is predicted by the hidden Markov model according to the vulnerability state transition probability to obtain the attack prediction path, and the vulnerability priority is obtained according to the order of the nodes on the attack prediction path.

6. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method for calculating data security according to the vulnerability priority and the vulnerability exploitability includes: Where l is the number of vulnerabilities, i is the i-th vulnerability, is the exploitability of the vulnerability, α i is the priority of the vulnerability, K(x i ,x) represents the network congestion rate of the channel where the vulnerability is located, and b is the bias coefficient.

7. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method for calculating the second evaluation score according to the data security and the compatibility includes: Obtaining a data security and compatibility importance scoring matrix, determining a membership function formed by a gradient distribution, and obtaining the data security and compatibility membership levels based on the membership function and the scoring matrix; The probability hesitant fuzzy average operator is used to aggregate each row of the rating matrix. The expression is: Among them, i=1,2,...,n represents the user, j=1,2 represents the data security index and compatibility index, represents the union operation of all membership levels, ξ is the number of users participating in the rating, and p ij is the score of the i-th user on the j-th indicator, represents the membership level of the i-th user to the j-th indicator, It means calculating the cumulative probability of each membership level. represents the product of all user membership levels, Indicates the consistency of ratings among all users; The aggregation results are normalized to obtain the index weights of compatibility and data security. The second evaluation score is calculated based on the index weights of compatibility and data security. The calculation formula is: f2=C·W i +E2·W j Among them, C is compatibility, E2 is data security, W is i is the compatibility weight, W j is the weight of data security.

8. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method of dividing the first evaluation score and the second evaluation score into quadrants to construct a quadrant evaluation system includes: Collect historical smart home parameters and communication data, calculate the historical first evaluation score and the historical second evaluation score, sort them from large to small, and select the lowest score of the top 70% of the historical first evaluation score and the top 80% of the historical second evaluation score as the dividing threshold points of the first evaluation score and the second evaluation score, and obtain V1 and V2 respectively; Construct a quadrant assessment system, including: The data points with the first evaluation score greater than or equal to V1 and the second evaluation score greater than or equal to V2 are regarded as points in the first quadrant area, and the evaluation level is excellent; The data points with the first evaluation score greater than or equal to V1 and the second evaluation score less than V2 are regarded as points in the second quadrant area, and the evaluation grade is good; The data points with the first evaluation score less than V1 and the second evaluation score greater than or equal to V2 are regarded as points in the third quadrant area, and the evaluation level is general; Data points whose first evaluation score is less than V1 and whose second evaluation score is less than V2 are regarded as points in the fourth quadrant area, and their evaluation level is poor.

9. The method for constructing a smart home quality risk assessment model based on big data according to claim 1, characterized in that: The method for constructing a quality risk assessment model based on a neural network includes: defining a neural network architecture, including an input layer, a hidden layer, and an output layer, wherein the input layer is used to obtain parameter data and communication data of a smart home, the hidden layer includes two logic processing layers, wherein the first logic processing layer is used to calculate a first evaluation score and a second evaluation score, the second logic processing layer is a threshold comparison layer, and the first evaluation score and the second evaluation score calculated by the first logic processing layer are used to obtain an evaluation grade based on a quadrant evaluation system, and the output layer includes two sublayers, wherein the first sublayer is used to output the evaluation grade obtained based on the second logic processing layer, and the second sublayer is a forward propagation layer used to output the probability distribution of each quadrant obtained by processing the first logic layer using a softmax function.

Images