An environment noise prediction method and system based on an improved neural network

Abstract

The present application relates to the technical field of data mining, and particularly relates to an environmental noise prediction method and system based on an improved neural network, which first extracts environmental noise data, divides training data sets and test data sets, randomly generates a population, obtains a current repulsion factor based on heuristic factors, fuses the current repulsion factor into a sine cosine operation to generate a new individual, and adaptively updates the repulsion factor, applies the improved neural network to predict environmental noise, uses a sine cosine algorithm based on a repulsion mechanism to optimize and design parameters of a neural network model for predicting environmental noise, and uses adaptive repulsion factors to promote individuals in the population to move away from poor individuals, thereby enhancing the search efficiency of the algorithm and improving the accuracy of the neural network model for predicting environmental noise, and the environmental noise prediction model obtained by training can improve the prediction accuracy of environmental noise.

Description

Technical Field

[0001] This invention relates to the field of data mining technology, and specifically to an environmental noise prediction method and system based on an improved neural network. Background Technology

[0002] As living standards improve, people are paying more and more attention to the comfort of their living environment. Environmental noise is a significant factor affecting this comfort. To better monitor environmental noise, many cities have installed environmental noise monitoring equipment in various locations. This equipment allows for convenient and real-time collection of environmental noise data from different areas. With the accumulation of environmental noise data, researchers are attempting to use data mining techniques to analyze this data, aiming to gain a comprehensive understanding of environmental noise pollution in different urban areas from multiple perspectives.

[0003] Environmental noise prediction is a crucial technique in environmental noise data analysis. To predict environmental noise, researchers need to analyze historical environmental noise data for a region to understand its changing patterns, and then use these patterns to predict future environmental noise pollution levels. Neural networks, an artificial intelligence method that simulates the human brain's information processing, have achieved some success in many engineering prediction applications. However, traditional neural networks, when applied to environmental noise prediction, are prone to getting trapped in local optima and suffer from insufficient prediction accuracy. Summary of the Invention

[0004] The purpose of this invention is to provide an environmental noise prediction method and system based on an improved neural network, which to a certain extent overcomes the shortcomings of traditional neural networks in predicting environmental noise, such as getting trapped in local optima and insufficient prediction accuracy, thereby improving the prediction accuracy of environmental noise.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide an environmental noise prediction method based on an improved neural network, the method comprising the following steps:

[0007] Step 1: Obtain the environmental noise dataset;

[0008] Step 2, input the population size NSize and the maximum search generations MaxG;

[0009] Step 3: Determine the neural network model for predicting environmental noise, and set the number of optimization design parameters ND for the neural network model;

[0010] Step 4: Set the current search algebra t = 0;

[0011] Step 5: Randomly generate the population NPop = {NA1, NA2, ..., NA} ni ,...,NA NSize}, where NA ni This represents the ni-th individual in the population, with individual indices ni = 1, 2, ..., NSize; individual NA... ni It stores ND optimization design parameters from the neural network model;

[0012] Step 6: Calculate the fitness value of each individual in the population NPop based on the environmental noise dataset. The individual with the smallest fitness value in the population NPop is recorded as the best individual BNA, and the individual with the largest fitness value in the population NPop is recorded as the worst individual WNA.

[0013] Step 7, set the rejection factor EF ni , where EF ni = rand(0,1), where rand represents a function to generate random real numbers;

[0014] Step 8, calculate the heuristic factor IF according to formula (1):

[0015]

[0016] Where sin represents the sine function; cos represents the cosine function; π represents pi; and kc represents the decay center.

[0017] Step 9: Calculate the current rejection factor CF according to formula (2). ni :

[0018]

[0019] Where pcr is a random real number between [0,1]; RW is a random real number between [0,0.5];

[0020] Step 10: Randomly generate a real number pes between [0,1].

[0021] Step 11: If PES is less than 0.5, proceed to step 12; otherwise, proceed to step 13.

[0022] Step 12: Perform a sinusoidal operation based on the repulsion mechanism according to formula (3) to generate a new individual NU. ni Then, proceed to step 14;

[0023] NU ni =NA ni +NK1×sin(NK2)×|NK3×BNA-NA ni |-CF ni ×(WNA-NAni (3);

[0024] in, NK2 is a random real number between [0, 2×π]; NK3 is a random real number between [0, 2].

[0025] Step 13: Perform a cosine operation based on the repulsion mechanism according to formula (4) to generate a new individual NU. ni Then, proceed to step 14;

[0026] NU ni =NA ni +NK1×cos(NK2)×|NK3×BNA-NA ni |-CF ni ×(WNA-NA ni (4);

[0027] Step 14: Calculate the new individual NU based on the environmental noise dataset. ni fitness value;

[0028] Step 15, determine the new individual NU ni Is the fitness value less than the individual's NA? ni If the fitness value is positive, then the new individual NU is used in the population. ni Replace individual NA ni Otherwise, maintain individual NA in the population. ni constant;

[0029] Step 16, update the rejection factor EF according to formula (5). ni :

[0030]

[0031] Step 17: Find the individual with the smallest fitness value from the population obtained in Step 15, and update the best individual BNA accordingly. Find the individual with the largest fitness value from the population obtained in Step 15, and update the worst individual WNA accordingly.

[0032] Step 18, set the current search algebra t = t + 1;

[0033] Step 19: If the current search algebra t is greater than MaxG, proceed to step 20; otherwise, proceed to step 8.

[0034] Step 20: Extract ND optimization design parameters of the neural network model from the optimal individual BNA, and use the obtained ND optimization design parameters to construct an environmental noise prediction model;

[0035] Step 21: Use the environmental noise prediction model to predict environmental noise.

[0036] Furthermore, the environmental noise dataset is determined in the following way:

[0037] Step 101: Extract environmental noise data from the environmental monitoring system;

[0038] Step 102: Preprocess the extracted environmental noise data to obtain an environmental noise dataset.

[0039] Furthermore, the neural network model used to predict environmental noise is a three-layer perceptron neural network.

[0040] Furthermore, the fitness value is calculated as follows:

[0041] For the ni-th individual NA in the population ni From individual NA ni Extract the ND optimization design parameters of its stored neural network model, and use the obtained ND optimization design parameters to construct an individual NA. ni Neural network model NModel ni Calculate individual NA ni Neural network model NModel ni Mean Square Error (NER) on Environmental Noise Dataset ni , will individual NA ni The fitness value is set to NER ni ; wherein, the individual NA ni Neural network model NModel ni The input variable is the equivalent sound level in each hour of the first time period before the current moment; the individual NA ni Neural network model NModel ni The output variable is the equivalent sound level for each hour in the second time period after the current time.

[0042] In a second aspect, embodiments of the present invention also provide a computer-readable storage medium storing an environmental noise prediction program based on an improved neural network, wherein when the environmental noise prediction program based on the improved neural network is executed by a processor, it implements the steps of the environmental noise prediction method based on the improved neural network as described in the first aspect.

[0043] Thirdly, embodiments of the present invention also provide an environmental noise prediction system based on an improved neural network, the system comprising:

[0044] At least one processor;

[0045] At least one memory for storing at least one program;

[0046] When the at least one program is executed by the at least one processor, the at least one processor implements the environmental noise prediction method based on the improved neural network described in the first aspect.

[0047] The beneficial effects of this invention are as follows: This invention discloses an environmental noise prediction method and system based on an improved neural network. This invention utilizes a sine / cosine algorithm based on a repulsion mechanism to optimize the parameters of the neural network model for predicting environmental noise. In the sine / cosine algorithm based on the repulsion mechanism, an adaptive repulsion factor is designed and integrated into the sine / cosine operations, enabling individuals in the population to adaptively distance themselves from weaker individuals, enhancing the algorithm's search efficiency, and thus improving the accuracy of the neural network model for predicting environmental noise. The environmental noise prediction model trained by this invention can improve the prediction accuracy of environmental noise. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a schematic diagram illustrating how the heuristic factor changes with the search algebra in an embodiment of the present invention. Detailed Implementation

[0050] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present invention can be combined with each other.

[0051] This invention provides an environmental noise prediction method based on an improved neural network, the method comprising the following steps:

[0052] Step 1: Obtain the environmental noise dataset;

[0053] The environmental noise dataset is determined in the following way:

[0054] Step 101: Extract environmental noise data from the environmental monitoring system; wherein the environmental noise data includes the equivalent sound level per hour;

[0055] Step 102: Preprocess the extracted environmental noise data to obtain an environmental noise dataset.

[0056] Step 2, input the population size NSize = 100 and the maximum search generations MaxG = 2000;

[0057] Step 3: Determine the neural network model for predicting environmental noise, and set the number of optimization design parameters ND for the neural network model;

[0058] In one embodiment, the neural network model used to predict environmental noise is set to a three-layer perceptron neural network with a 10-15-1 structure, and the number of optimization design parameters of the neural network model is set to ND = 181.

[0059] Step 4: Set the current search algebra t = 0;

[0060] Step 5: Randomly generate the population NPop = {NA1, NA2, ..., NA} ni ,...,NA NSize}, where NA ni This represents the ni-th individual in the population, with individual indices ni = 1, 2, ..., NSize; individual NA... ni Each individual in the population NPop stores ND optimization design parameters from the neural network model; that is, each individual in the population NPop stores ND optimization design parameters from the neural network model.

[0061] Step 6: Calculate the fitness value of each individual in the population NPop based on the environmental noise dataset; find the individual with the smallest fitness value in the population NPop and denote it as the best individual BNA; find the individual with the largest fitness value in the population NPop and denote it as the worst individual WNA.

[0062] In a preferred embodiment, the fitness value is calculated as follows:

[0063] For the ni-th individual NA in the population ni From individual NA ni Extract the ND optimization design parameters of its stored neural network model, and use the obtained ND optimization design parameters to construct an individual NA. ni Neural network model NModel ni Calculate individual NA ni Neural network model NModel ni Mean Square Error (NER) on Environmental Noise Dataset ni , will individual NA ni The fitness value is set to NER ni ; wherein, the individual NA ni Neural network model NModel ni The input variable is the equivalent sound level in each hour of the first time period before the current moment; the individual NA niNeural network model NModel ni The output variable is the equivalent sound level for each hour in the second time period after the current time.

[0064] In one embodiment, the individual NA ni Neural network model NModel ni The input variable is set to the equivalent sound level of each hour in the previous 10 hours; the individual NA ni Neural network model NModel ni The output variable is set to the equivalent sound level within one hour after the current time.

[0065] Step 7, set the rejection factor EF ni , where EF ni =rand(0,1), where rand represents a function to generate random real numbers; that is, EF ni A real number randomly generated from (0,1);

[0066] Step 8: Calculate the heuristic factor IF according to formula (1):

[0067]

[0068] Where sin represents the sine function; cos represents the cosine function; π represents pi; kc represents the decay center; and the result of a single calculation of the heuristic factor IF is as follows: Figure 1 As shown;

[0069] Step 9: Calculate the current rejection factor CF according to formula (2). ni :

[0070]

[0071] Where pcr is a random real number between [0,1]; RW is a random real number between [0,0.5];

[0072] Step 10: Randomly generate a real number pes between [0,1].

[0073] Step 11: If PES is less than 0.5, proceed to step 12; otherwise, proceed to step 13.

[0074] Step 12: Perform a sinusoidal operation based on the repulsion mechanism according to formula (3) to generate a new individual NU. ni Then, proceed to step 14;

[0075] NU ni =NA ni +NK1×sin(NK2)×|NK3×BNA-NA ni|-CF ni ×(WNA-NA ni (3);

[0076] in, NK2 is a random real number between [0, 2×π]; NK3 is a random real number between [0, 2].

[0077] Step 13: Perform a cosine operation based on the repulsion mechanism according to formula (4) to generate a new individual NU. ni Then, proceed to step 14;

[0078] NU ni =NA ni +NK1×cos(NK2)×|NK3×BNA-NA ni |-CF ni ×(WNA-NA ni (4);

[0079] Step 14: Calculate the new individual NU based on the environmental noise dataset. ni fitness value;

[0080] Step 15, determine the new individual NU ni Is the fitness value less than the individual's NA? ni If the fitness value is positive, then the new individual NU is used in the population. ni Replace individual NA ni Otherwise, maintain individual NA in the population. ni constant;

[0081] Step 16, update the rejection factor EF according to formula (5). ni :

[0082]

[0083] Step 17: Find the individual with the smallest fitness value from the population obtained in Step 15, and update the best individual BNA accordingly. Find the individual with the largest fitness value from the population obtained in Step 15, and update the worst individual WNA accordingly.

[0084] That is, the optimal individual BNA determined in step 6 is updated to the individual with the smallest fitness value in the population obtained in step 15, and the worst individual WNA determined in step 6 is updated to the individual with the largest fitness value in the population obtained in step 15.

[0085] Step 18, set the current search algebra t = t + 1;

[0086] Step 19: If the current search algebra t is greater than MaxG, proceed to step 20; otherwise, proceed to step 8.

[0087] Step 20: Extract ND optimization design parameters of the neural network model from the optimal individual BNA, and use the obtained ND optimization design parameters to construct an environmental noise prediction model;

[0088] Step 21: Use the environmental noise prediction model to predict environmental noise.

[0089] Corresponding to the above method embodiments, the present invention also provides a computer-readable storage medium storing an environmental noise prediction program based on an improved neural network. When the environmental noise prediction program based on the improved neural network is executed by a processor, it implements the steps of the environmental noise prediction method based on the improved neural network as described in any of the above embodiments.

[0090] Corresponding to the above method embodiments, this invention also provides an environmental noise prediction system based on an improved neural network, the system comprising:

[0091] At least one processor;

[0092] At least one memory for storing at least one program;

[0093] When the at least one program is executed by the at least one processor, the at least one processor implements the environmental noise prediction method based on the improved neural network described in any of the above embodiments.

[0094] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0095] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the environmental noise prediction system based on the improved neural network, connecting various parts of the entire operable device of the environmental noise prediction system based on the improved neural network through various interfaces and lines.

[0096] The memory can be used to store the computer program and / or modules. The processor implements various functions of the environmental noise prediction system based on the improved neural network by running or executing the computer program and / or modules stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart-media-card (SMC), secure-digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0097] Although the invention has been described in considerable detail and particularly with regard to several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, but should be regarded as effectively covering the intended scope of the invention by referring to the appended claims and taking into account the broad possible interpretations of these claims provided by the prior art. Furthermore, the invention has been described above with regard to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

Claims

1. An environmental noise prediction method based on an improved neural network, characterized in that, The method includes the following steps: Step 1: Obtain the environmental noise dataset; Step 2, input the population size NSize and the maximum search generations MaxG; Step 3: Determine the neural network model for predicting environmental noise, and set the number of optimization design parameters ND for the neural network model; Step 4: Set the current search algebra t = 0; Step 5: Randomly generate the population NPop = {NA1, NA2, ..., NA} ni ,...,NA NSize }, where NA ni This represents the ni-th individual in the population, with individual indices ni = 1, 2, ..., NSize; individual NA... ni It stores ND optimization design parameters from the neural network model; Step 6: Calculate the fitness value of each individual in the population NPop based on the environmental noise dataset. The individual with the smallest fitness value in the population NPop is recorded as the best individual BNA, and the individual with the largest fitness value in the population NPop is recorded as the worst individual WNA. Step 7, set the rejection factor EF ni , where EF ni = rand(0,1), where rand represents a function to generate random real numbers; Step 8, calculate the heuristic factor IF according to formula (1): Where sin represents the sine function; cos represents the cosine function; π represents pi; and kc represents the decay center. Step 9: Calculate the current rejection factor CF according to formula (2). ni : Where pcr is a random real number between [0,1]; RW is a random real number between [0,0.5]; Step 10: Randomly generate a real number pes between [0,1]. Step 11: If PES is less than 0.5, proceed to step 12; otherwise, proceed to step 13. Step 12: Perform a sinusoidal operation based on the repulsion mechanism according to formula (3) to generate a new individual NU. ni Then, proceed to step 14; NU ni JNA ni +NK1×sin(NK2)×|NK3×BNA-NA ni |-CF ni ×(WNA-NA ni ) (3) in, NK2 is a random real number between [0, 2×π]; NK3 is a random real number between [0, 2]. Step 13: Perform a cosine operation based on the repulsion mechanism according to formula (4) to generate a new individual NU. ni Then, proceed to step 14; NU ni JNA ni +NK1×cos(NK2)×|NK3×BNA-NA ni |-CF ni ×(WNA-NA ni ) (4) Step 14: Calculate the new individual NU based on the environmental noise dataset. ni fitness value; Step 15, determine the new individual NU ni Is the fitness value less than the individual's NA? ni If the fitness value is positive, then the new individual NU is used in the population. ni Replace individual NA ni Otherwise, maintain individual NA in the population. ni constant; Step 16, update the rejection factor EF according to formula (5). ni : Step 17: Find the individual with the smallest fitness value from the population obtained in Step 15, and update the best individual BNA accordingly. Find the individual with the largest fitness value from the population obtained in Step 15, and update the worst individual WNA accordingly. Step 18, set the current search algebra t = t + 1; Step 19: If the current search algebra t is greater than MaxG, proceed to step 20; otherwise, proceed to step 8. Step 20: Extract ND optimization design parameters of the neural network model from the optimal individual BNA, and use the obtained ND optimization design parameters to construct an environmental noise prediction model; Step 21: Use the environmental noise prediction model to predict environmental noise.

2. The environmental noise prediction method based on an improved neural network according to claim 1, characterized in that, The environmental noise dataset was determined in the following way: Step 101: Extract environmental noise data from the environmental monitoring system; Step 102: Preprocess the extracted environmental noise data to obtain an environmental noise dataset.

3. The environmental noise prediction method based on an improved neural network according to claim 1, characterized in that, The neural network model used to predict environmental noise is a three-layer perceptron neural network.

4. The environmental noise prediction method based on an improved neural network according to claim 3, characterized in that, The fitness value is calculated as follows: For the ni-th individual NA in the population ni From individual NA ni Extract the ND optimization design parameters of its stored neural network model, and use the obtained ND optimization design parameters to construct an individual NA. ni Neural network model NModel ni Calculate individual NA ni Neural network model NModel ni Mean Square Error (NER) on Environmental Noise Dataset ni , will individual NA ni The fitness value is set to NER ni ; wherein, the individual NA ni Neural network model NModel ni The input variable is the equivalent sound level in each hour of the first time period before the current moment; the individual NA ni Neural network model NModel ni The output variable is the equivalent sound level for each hour in the second time period after the current time.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the environmental noise prediction method based on an improved neural network as described in any one of claims 1 to 4.

6. An environmental noise prediction system based on an improved neural network, characterized in that, The system includes: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the environmental noise prediction method based on the improved neural network as described in any one of claims 1 to 4.

Images