Mobile phone mainboard performance parameter calibration method based on artificial intelligence

Abstract

The invention discloses a mobile phone mainboard performance parameter calibration method based on artificial intelligence, and belongs to the technical field of mobile phone mainboards. According to the method, the problem that the performance parameters of the mobile phone mainboard cannot be calibrated in the prior art is solved, the artificial intelligence model is trained through the training set data, the trained artificial intelligence model is evaluated and optimized, the prediction accuracy of the artificial intelligence model can be improved, and the performance of the mobile phone mainboard can be calibrated. The optimized artificial intelligence model is applied to calibration of the performance parameters of the mobile phone mainboard, so that the artificial intelligence model can predict the performance of the mobile phone mainboard, the difference between the performance parameters, predicted by the artificial intelligence model, of the mobile phone mainboard and the actual performance can be helped, then optimization suggestions of the mobile phone mainboard are pointed out according to the prediction result, and the optimization efficiency of the mobile phone mainboard is improved. According to the method, the mobile phone mainboard can be calibrated, then the parameters of the mobile phone mainboard are adjusted to optimize the performance of the mobile phone mainboard, finally, the calibration effect of the performance parameters of the mobile phone mainboard can be verified through testing, and the calibration accuracy of the performance parameters of the mobile phone mainboard is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile phone motherboards, and in particular to a method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence. Background Art

[0002] The mobile phone motherboard, also known as the main board or system board, is one of the core components inside the mobile phone. It is like a "central hub" that connects various components inside the mobile phone, such as the CPU, memory, storage chips, communication modules, etc., to ensure that they can work together. The performance and stability of the mobile phone motherboard directly determine the overall performance of the mobile phone. An excellent motherboard design can ensure that the mobile phone remains stable while running at high speed, extending the service life of the mobile phone. However, the existing mobile phone motherboard cannot be calibrated for performance parameters. If the performance parameters of the mobile phone motherboard cannot be calibrated, the communication quality will be reduced and the equipment cost will be increased, and it will not be possible to ensure that every mobile phone shipped can meet the preset performance standards, and it will not provide a better user experience. At the same time, it will increase costs and fail to achieve economic benefits; therefore, it does not meet the existing needs. For this, we propose a method for calibrating the performance parameters of mobile phone motherboards based on artificial intelligence. Summary of the invention

[0003] The purpose of the present invention is to provide a method for calibrating the performance parameters of a mobile phone motherboard based on artificial intelligence, by training an artificial intelligence model, and evaluating and optimizing the trained artificial intelligence model to improve the accuracy of the artificial intelligence model's prediction, and applying the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone motherboard, and by acquiring and inputting the current status data of the mobile phone motherboard, the artificial intelligence model is enabled to predict the performance of the mobile phone motherboard, thereby helping to find the difference between the performance parameters of the mobile phone motherboard predicted by the artificial intelligence model and the actual performance, and then pointing out the optimization suggestions for the mobile phone motherboard based on the prediction results, and then adjusting the parameters of the mobile phone motherboard to optimize the performance of the mobile phone motherboard, and verifying the effect of the calibration of the performance parameters of the mobile phone motherboard through testing, thereby solving the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence, comprising the following steps:

[0005] Step 1: Collect relevant performance data of mobile phone motherboards, and at the same time, establish an artificial intelligence model to pre-process the collected performance data of mobile phone motherboards and integrate them into training set data;

[0006] Step 2: Use the training set data to train the artificial intelligence model, so that the trained artificial intelligence model can predict the performance of the mobile phone motherboard;

[0007] Step 3: After training is completed, the AI ​​model is evaluated and optimized based on the evaluation results;

[0008] Step 4: Apply the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone motherboard to achieve the calibration of the performance parameters of the mobile phone motherboard;

[0009] Step 5: Based on the prediction of the artificial intelligence model, adjust the parameters of the mobile phone motherboard to optimize the performance of the mobile phone motherboard;

[0010] Step 6: Test the mobile phone motherboard after parameter adjustment to verify the effect of the mobile phone motherboard performance parameter calibration through testing.

[0011] Furthermore, the step 1 specifically includes the following steps:

[0012] Data collection: collect relevant performance data of the mobile phone motherboard through special sensors or test tools. The collected performance data is the basis of model training;

[0013] Model building: Building artificial intelligence models based on the collected performance data;

[0014] Data preprocessing: preprocess the collected performance data, including data cleaning, data conversion and data reduction;

[0015] Data integration: The pre-processed performance data is integrated into training set data, which is used to train the established artificial intelligence model.

[0016] Furthermore, the preprocessing includes:

[0017] Data cleaning: remove errors, missing values, and noise from performance data;

[0018] Data conversion: converting performance data into a unified data format or data type;

[0019] Data reduction: Reduce the dimensionality of performance data through feature extraction, reduce computational costs and improve AI model performance.

[0020] Furthermore, the preprocessing further includes:

[0021] Extract the data information before removing the noise;

[0022] According to the value of the data information before the noise is removed, obtaining a data distribution diagram of the data information before the noise is removed as a first data distribution diagram;

[0023] According to the value of the data information after the noise is removed, a data distribution diagram of the data information after the noise is removed is obtained as a second data distribution diagram;

[0024] Compare the first data distribution graph and the second data distribution graph to obtain a similarity value between the first data distribution graph and the second data distribution graph;

[0025] Comparing the similarity value between the first data distribution graph and the second data distribution graph with a preset similarity threshold;

[0026] When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, performing data quality evaluation on the data information after noise removal;

[0027] When the similarity value between the first data distribution graph and the second data distribution graph is lower than a preset similarity threshold, traversing each data information of the first data distribution graph and each data information of the second data distribution graph, and obtaining the distribution position of each data information in the first data distribution graph and the second data distribution graph;

[0028] Obtaining a distribution position straight-line distance according to the distribution position of each piece of data information in the first data distribution graph and the second data distribution graph;

[0029] The data distortion coefficient is obtained according to the straight-line distance corresponding to the distribution position of each data information in the first data distribution diagram and the second data distribution diagram; wherein the data distortion coefficient is obtained by the following formula:

[0030]

[0031] Wherein, S represents the data distortion coefficient; k represents the number of data information contained in the first data distribution diagram and the second data distribution diagram; L i Indicates the straight-line distance of the distribution location corresponding to the i-th data information; L yi represents the minimum straight-line distance allowed for the distribution position corresponding to the ith data information that meets the similarity threshold; m represents the number of data information whose straight-line distances of the distribution positions in the k data information exceed the minimum straight-line distance allowed for the corresponding distribution positions; n represents the number of data information whose straight-line distances of the distribution positions in the k data information are 0; q represents the number of data information whose straight-line distances of the distribution positions in the k data information do not exceed the minimum straight-line distance allowed for the corresponding distribution positions, and whose straight-line distances are non-zero; s 01 、s 02 and 03 Respectively represent the first adjustment coefficient, the second adjustment coefficient and the third adjustment coefficient;

[0032] The first adjustment coefficient is obtained by the following formula:

[0033]

[0034] Among them, s 01 represents the first adjustment coefficient; m represents the number of data information whose straight-line distance of the distribution position among the k data information exceeds the minimum straight-line distance allowed by the corresponding distribution position; L 01i Indicates the straight-line distance corresponding to the data information whose straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed by its corresponding distribution position; L y01i The minimum straight-line distance allowed for the distribution position corresponding to the data information indicating that the straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed for the corresponding distribution position;

[0035] At the same time, the second adjustment coefficient is obtained by the following formula:

[0036]

[0037] Among them, s 02 represents the second adjustment coefficient; n represents the number of data information whose straight-line distance of distribution position is 0 among the k data information; L 02xi and L 02hi They respectively represent the straight-line distance corresponding to the previous data information and the straight-line distance corresponding to the next data information adjacent to the data information whose straight-line distance is 0 at the i-th distribution position; yx02i and L yh02i They respectively represent the minimum straight-line distance allowed at the distribution position corresponding to the previous data information and the minimum straight-line distance allowed at the distribution position corresponding to the next data information;

[0038] At the same time, the third adjustment coefficient is obtained by the following formula:

[0039]

[0040] Among them, s 03 represents the third adjustment coefficient; q represents the number of data information whose straight-line distance of the distribution position in the k data information does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and whose straight-line distance is non-zero; L 03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by its corresponding distribution position, and the straight-line distance corresponds to the data information with non-zero straight-line distance; L y03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and the minimum straight-line distance allowed by the distribution position corresponding to the data information with non-zero straight-line distance;

[0041] When the data distortion coefficient exceeds a preset coefficient threshold, an abnormal denoising operation alarm is issued;

[0042] When the data distortion coefficient does not exceed a preset coefficient threshold, the data information before noise removal is re-cleaned.

[0043] Furthermore, when the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, a data quality evaluation is performed on the data information after the noise is removed, including:

[0044] When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, extracting the signal-to-noise ratio, data integrity and noise residue of the data information after the noise is removed;

[0045] The data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by using the signal-to-noise ratio, data integrity and noise residual of the data information after the noise is removed; wherein the data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by the following formula:

[0046]

[0047] Where W represents the data quality evaluation parameter corresponding to the data information after noise removal; p represents the data integrity of the i-th data information; R i represents the noise residual corresponding to the i-th data information; C i represents the maximum allowable noise residual corresponding to the i-th data information; B i represents the signal-to-noise ratio of the data information after noise removal corresponding to the i-th data information; B xi represents the signal-to-noise ratio of the data information before noise removal corresponding to the i-th data information;

[0048] Comparing the data quality evaluation parameter corresponding to the data information after noise removal with a preset data quality evaluation parameter threshold;

[0049] When the data quality evaluation parameter corresponding to the data information after noise removal is lower than a preset data quality evaluation parameter threshold, a data quality abnormality alarm is issued.

[0050] Furthermore, the step 2 specifically includes the following steps:

[0051] Substitute the integrated training set data into the artificial intelligence model for training and learning;

[0052] Artificial intelligence models are based on deep learning algorithms or machine learning algorithms to learn patterns and features in the training set data;

[0053] By learning patterns and features from the training set data, the AI ​​model can predict the performance of the mobile phone motherboard.

[0054] Furthermore, the step three specifically includes the following steps:

[0055] Performance evaluation: Perform performance evaluation on the AI ​​model after training. The evaluation results will be used to optimize and improve the AI ​​model;

[0056] Among them, the performance evaluation uses the cross-validation method to divide the training set data into subsets to evaluate the artificial intelligence model;

[0057] Among them, the cross-validation method includes:

[0058] Data set partitioning: Divide the training set data into multiple non-overlapping subsets;

[0059] Model evaluation: Based on the cross-validation method, multiple non-overlapping subsets of data are substituted into the artificial intelligence model, and the artificial intelligence model is repeatedly evaluated using different subsets of data to finally obtain the evaluation results;

[0060] Result analysis: Analyze the evaluation results to find out the factors that affect the performance of the AI ​​model, and then optimize and improve the AI ​​model;

[0061] Model optimization and improvement: Based on the evaluation results of the artificial intelligence model, the artificial intelligence model is optimized and improved. The optimized artificial intelligence model will be used in the calibration of the performance parameters of the mobile phone motherboard in step four.

[0062] Furthermore, the performance evaluation of the AI ​​model includes the evaluation of the accuracy, stability, scalability and reliability of the AI ​​model, and the optimization and improvement of the AI ​​model includes improving the training method and adding or replacing features.

[0063] Furthermore, the step 4 specifically includes the following steps:

[0064] Apply the optimized AI model to calibrate the performance parameters of mobile phone motherboards;

[0065] Obtain and input the status data of the current mobile phone motherboard, and the artificial intelligence model predicts the performance of the mobile phone motherboard;

[0066] Based on the prediction results, point out the optimization suggestions for the mobile phone motherboard;

[0067] Calibrate the performance parameters of the mobile phone motherboard according to the existing optimization suggestions and determine the parameters that need to be adjusted.

[0068] Furthermore, the step six specifically includes the following steps:

[0069] After the mobile phone mainboard has been parameter adjusted in step 5, actual data of the mobile phone mainboard is obtained;

[0070] Integrate the acquired actual data into test set data;

[0071] Substitute the integrated test set data into the artificial intelligence model for verification;

[0072] The test results are used to determine how well the AI ​​model’s predictions match the actual data.

[0073] Compared with the prior art, the present invention has the following beneficial effects:

[0074] The present invention collects relevant performance data of a mobile phone mainboard, pre-processes the collected performance data and integrates them into training set data, establishes an artificial intelligence model, substitutes the training set data into the artificial intelligence model for training and learning, trains the artificial intelligence model, and evaluates and optimizes the trained artificial intelligence model, thereby improving the accuracy of the artificial intelligence model prediction, and applies the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone mainboard, and enables the artificial intelligence model to predict the performance of the mobile phone mainboard by acquiring and inputting the status data of the current mobile phone mainboard, thereby helping to find the difference between the performance parameters of the mobile phone mainboard predicted by the artificial intelligence model and the actual performance, and then points out the optimization suggestions for the mobile phone mainboard according to the prediction results, and then adjusts the parameters of the mobile phone mainboard to optimize the performance of the mobile phone mainboard, and finally tests the mobile phone mainboard after parameter adjustment, and verifies the effect of the calibration of the performance parameters of the mobile phone mainboard through the test, thereby ensuring the accuracy of the calibration of the performance parameters of the mobile phone mainboard. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 It is a flow chart of the method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence of the present invention;

[0076] Figure 2 A schematic diagram of the process of data collection and model building of the present invention;

[0077] Figure 3 A schematic diagram of the process of training the artificial intelligence model of the present invention;

[0078] Figure 4 A schematic diagram of the process of evaluating and optimizing the artificial intelligence model of the present invention;

[0079] Figure 5 A schematic diagram of a process for calibrating performance parameters of a mobile phone motherboard according to the present invention;

[0080] Figure 6 The present invention is a flowchart of verifying the performance parameter calibration effect of a mobile phone motherboard. DETAILED DESCRIPTION

[0081] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0082] In order to solve the technical problem that the existing mobile phone motherboard cannot be calibrated for performance parameters, the performance parameters of the mobile phone motherboard cannot be calibrated, which will reduce the communication quality and increase the equipment cost, and thus cannot ensure that every mobile phone shipped can meet the preset performance standards, cannot provide a better user experience, and at the same time increase the cost and cannot achieve economic benefits, please refer to Figure 1-Figure 6 , this embodiment provides the following technical solutions:

[0083] A method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence, comprising the following steps:

[0084] Step 1: Collect relevant performance data of mobile phone motherboards, and at the same time, establish an artificial intelligence model to pre-process the collected performance data of mobile phone motherboards and integrate them into training set data;

[0085] Step 2: Use the training set data to train the artificial intelligence model, so that the trained artificial intelligence model can predict the performance of the mobile phone motherboard;

[0086] Step 3: After training is completed, the AI ​​model is evaluated and optimized based on the evaluation results;

[0087] Step 4: Apply the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone motherboard to achieve the calibration of the performance parameters of the mobile phone motherboard;

[0088] Step 5: Based on the prediction of the artificial intelligence model, adjust the parameters of the mobile phone motherboard to optimize the performance of the mobile phone motherboard;

[0089] Step 6: Test the mobile phone motherboard after parameter adjustment to verify the effect of the mobile phone motherboard performance parameter calibration through testing.

[0090] The technical effect of the above content is: the method for calibrating the performance parameters of mobile phone motherboards based on artificial intelligence is composed of collecting data and establishing models, training models, evaluating and optimizing models, model application, parameter adjustment and calibration effect verification. Through the above process, the performance parameters of the mobile phone motherboard can be calibrated and the calibration effect can be tested and verified, thereby improving communication quality and reducing equipment costs, thereby ensuring that every mobile phone leaving the factory can meet the preset performance standards.

[0091] Step 1 specifically includes the following steps:

[0092] Data collection: collect relevant performance data of the mobile phone motherboard through special sensors or test tools. The collected performance data is the basis of model training;

[0093] Model building: Building artificial intelligence models based on the collected performance data;

[0094] Data preprocessing: preprocess the collected performance data, including data cleaning, data conversion and data reduction;

[0095] Data integration: The pre-processed performance data is integrated into training set data, which is used to train the established artificial intelligence model.

[0096] The technical effects of the above content are as follows: first, it is necessary to collect data for training the model and to establish an artificial intelligence model. Data collection is the basis of model training. The performance evaluation of the artificial intelligence model depends on high-quality data sets. For the calibration of the performance parameters of the mobile phone motherboard, it is necessary to first prepare relevant data sets. These data sets should contain the performance data of the mobile phone motherboard under different conditions, so that the changing patterns of performance parameters can be learned when training the artificial intelligence model. Data collection can be carried out through special sensors or test tools to collect relevant performance data of the mobile phone motherboard. The collected performance data will be integrated into training set data after data cleaning, data conversion, and data reduction. The integrated training set data can be used to train the established artificial intelligence model.

[0097] Pre-processing, including:

[0098] Data cleaning: remove errors, missing values, and noise from performance data;

[0099] Data conversion: converting performance data into a unified data format or data type;

[0100] Data reduction: Reduce the dimensionality of performance data through feature extraction, reduce computational costs and improve AI model performance.

[0101] The technical effect of the above content is: the main purpose of data cleaning is to improve the quality of performance data, reduce the probability of errors in the data statistics process, and ensure the accuracy and validity of performance data. Through data cleaning, the erroneous, inconsistent, incomplete and redundant data in the performance data can be removed or repaired to make the performance data more accurate, reliable and useful, thereby improving the accuracy and reliability of model predictions. The data cleaning method mainly includes the following steps: data review, a comprehensive review of the performance data, including checking the type of performance data, missing values, outliers, duplicate records, etc.; missing value processing, processing missing values, and you can choose to delete missing values; outlier detection and processing: identifying and processing outliers, you can detect and process outliers by setting thresholds; duplicate data processing, Identify and process duplicate data, you can choose to delete duplicate records or merge duplicate records; data verification, verify the performance data to ensure the accuracy and completeness of the performance data; the main purpose of data conversion is to ensure the consistency of performance data, through data format unification, to ensure the consistency of data format, through data type conversion, the performance data is converted into a unified data type, converting the performance data into a unified data format or data type can improve the security and maintainability of the performance data. Finally, the main purpose of data reduction is to reduce the dimension of performance data through feature extraction, thereby reducing computing costs and improving the performance of artificial intelligence models. Preprocessing the performance data through the above process can make the original data more standardized, accurate and usable, thereby improving the effect and accuracy of model prediction.

[0102] Specifically, the preprocessing further includes:

[0103] Extract the data information before removing the noise;

[0104] According to the value of the data information before the noise is removed, obtaining a data distribution diagram of the data information before the noise is removed as a first data distribution diagram;

[0105] According to the value of the data information after the noise is removed, a data distribution diagram of the data information after the noise is removed is obtained as a second data distribution diagram;

[0106] Compare the first data distribution graph and the second data distribution graph to obtain a similarity value between the first data distribution graph and the second data distribution graph;

[0107] Comparing the similarity value between the first data distribution graph and the second data distribution graph with a preset similarity threshold;

[0108] When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, performing data quality evaluation on the data information after noise removal;

[0109] When the similarity value between the first data distribution graph and the second data distribution graph is lower than a preset similarity threshold, traversing each data information of the first data distribution graph and each data information of the second data distribution graph, and obtaining the distribution position of each data information in the first data distribution graph and the second data distribution graph;

[0110] Obtaining a distribution position straight-line distance according to the distribution position of each piece of data information in the first data distribution graph and the second data distribution graph;

[0111] The data distortion coefficient is obtained according to the straight-line distance corresponding to the distribution position of each data information in the first data distribution diagram and the second data distribution diagram; wherein the data distortion coefficient is obtained by the following formula:

[0112]

[0113] Wherein, S represents the data distortion coefficient; k represents the number of data information contained in the first data distribution diagram and the second data distribution diagram; L i Indicates the straight-line distance of the distribution location corresponding to the i-th data information; L yi represents the minimum straight-line distance allowed for the distribution position corresponding to the ith data information that meets the similarity threshold; m represents the number of data information whose straight-line distances of the distribution positions in the k data information exceed the minimum straight-line distance allowed for the corresponding distribution positions; n represents the number of data information whose straight-line distances of the distribution positions in the k data information are 0; q represents the number of data information whose straight-line distances of the distribution positions in the k data information do not exceed the minimum straight-line distance allowed for the corresponding distribution positions, and whose straight-line distances are non-zero; s 01 、s 02 and 03 Respectively represent the first adjustment coefficient, the second adjustment coefficient and the third adjustment coefficient;

[0114] The first adjustment coefficient is obtained by the following formula:

[0115]

[0116] Among them, s 01 represents the first adjustment coefficient; m represents the number of data information whose straight-line distance of the distribution position among the k data information exceeds the minimum straight-line distance allowed by the corresponding distribution position; L 01i Indicates the straight-line distance corresponding to the data information whose straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed by its corresponding distribution position; L y01i The minimum straight-line distance allowed for the distribution position corresponding to the data information indicating that the straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed for the corresponding distribution position;

[0117] At the same time, the second adjustment coefficient is obtained by the following formula:

[0118]

[0119] Among them, s 02 represents the second adjustment coefficient; n represents the number of data information whose straight-line distance of distribution position is 0 among the k data information; L 02xi and L 02hi They respectively represent the straight-line distance corresponding to the previous data information and the straight-line distance corresponding to the next data information adjacent to the data information whose straight-line distance is 0 at the i-th distribution position; yx02i and L yh02i They respectively represent the minimum straight-line distance allowed at the distribution position corresponding to the previous data information and the minimum straight-line distance allowed at the distribution position corresponding to the next data information;

[0120] At the same time, the third adjustment coefficient is obtained by the following formula:

[0121]

[0122] Among them, s 03 represents the third adjustment coefficient; q represents the number of data information whose straight-line distance of the distribution position in the k data information does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and whose straight-line distance is non-zero; L 03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by its corresponding distribution position, and the straight-line distance corresponds to the data information with non-zero straight-line distance; L y03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and the minimum straight-line distance allowed by the distribution position corresponding to the data information with non-zero straight-line distance;

[0123] When the data distortion coefficient exceeds a preset coefficient threshold, an abnormal denoising operation alarm is issued;

[0124] When the data distortion coefficient does not exceed a preset coefficient threshold, the data information before noise removal is re-cleaned.

[0125] The technical effect of the above technical solution is: by comparing the data distribution graphs before and after noise removal (the first data distribution graph and the second data distribution graph) and calculating their similarity values, the technical solution first conducts a preliminary evaluation of the effect of data denoising. This method not only considers the overall distribution change of data values, but also measures the impact of denoising on the overall structure of the data through the quantitative index of the similarity value.

[0126] When the similarity value is lower than the preset threshold, it means that the denoising process may have introduced data distortion or failed to effectively remove noise. At this point, by further analyzing the position change (straight-line distance) of each data point in the two data distribution graphs, the degree and specific location of data distortion can be accurately identified. This method improves the accuracy of anomaly detection and helps to quickly locate the problem.

[0127] By introducing the data distortion coefficient S, the technical solution realizes the quantitative evaluation of the degree of data distortion. This coefficient takes into account multiple factors, including the number of data points exceeding the allowable straight-line distance, the number of data points with a straight-line distance of 0, and the number of data points with a straight-line distance that is not exceeded but is non-zero, and is weighted and adjusted through the first, second and third adjustment coefficients. This quantitative evaluation provides a reliable basis for subsequent abnormal processing. When the data distortion coefficient exceeds the preset threshold, the system can promptly issue a denoising operation abnormality alarm to remind users or system administrators to pay attention to and deal with possible denoising problems. This timeliness helps to reduce errors in subsequent analysis or decision-making caused by data distortion. When the data distortion degree does not exceed the preset threshold, the system chooses to re-clean the data information before noise removal. This flexibility ensures the continuity and effectiveness of data processing while avoiding unnecessary repeated denoising operations.

[0128] In summary, this technical solution effectively improves the quality and efficiency of data processing by comprehensively evaluating the data denoising effect, accurately detecting data distortion, quantitatively evaluating the degree of distortion, and promptly responding to exception processing.

[0129] Specifically, when the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, a data quality evaluation is performed on the data information after the noise is removed, including:

[0130] When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, extracting the signal-to-noise ratio, data integrity and noise residue of the data information after the noise is removed;

[0131] The data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by using the signal-to-noise ratio, data integrity and noise residual of the data information after the noise is removed; wherein the data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by the following formula:

[0132]

[0133] Where W represents the data quality evaluation parameter corresponding to the data information after noise removal; p represents the data integrity of the i-th data information; R i represents the noise residual corresponding to the i-th data information; Ci represents the maximum allowable noise residual corresponding to the i-th data information; B i represents the signal-to-noise ratio of the data information after noise removal corresponding to the i-th data information; B xi represents the signal-to-noise ratio of the data information before noise removal corresponding to the i-th data information;

[0134] Comparing the data quality evaluation parameter corresponding to the data information after noise removal with a preset data quality evaluation parameter threshold;

[0135] When the data quality evaluation parameter corresponding to the data information after noise removal is lower than a preset data quality evaluation parameter threshold, a data quality abnormality alarm is issued.

[0136] The technical effect of the above technical solution is: when the similarity between the data distribution map after noise removal and the original data distribution map meets the preset threshold, the technical solution further extracts multiple key quality indicators of the data information after noise removal, including signal-to-noise ratio, data integrity and noise residue. These indicators together constitute a comprehensive perspective for data quality evaluation, which can reflect the true status of the data in more detail. The data quality evaluation parameter W corresponding to the information of the data after noise removal is calculated by a formula. This parameter comprehensively considers the data integrity, the ratio of the noise residue to the maximum allowable value, and the change in the signal-to-noise ratio. This quantitative evaluation method makes the data quality have a comparable and measurable standard, which helps to objectively evaluate the effect of data processing.

[0137] The technical solution presets the threshold of the data quality evaluation parameter. By comparing it with the data quality evaluation parameter actually calculated, it can be flexibly judged whether the data quality meets the requirements. This mechanism provides flexible adjustment space for data quality control in different application scenarios. When the data quality evaluation parameter is lower than the preset threshold, the system can promptly issue an abnormal data quality alarm. This timeliness helps to promptly discover and deal with data quality problems, avoiding subsequent analysis or decision-making errors caused by poor data quality. Through the above technical solution, the data processing process is more strictly and comprehensively monitored and evaluated. This helps to ensure that the processed data has higher reliability and accuracy, thereby improving the scientificity and effectiveness of data analysis and decision-making.

[0138] In summary, this technical solution effectively improves the reliability of data processing and the control level of data quality through measures such as detailed data quality assessment, quantitative data quality evaluation parameters, flexible threshold setting, and timely data quality anomaly alarm.

[0139] Step 2 specifically includes the following steps:

[0140] Substitute the integrated training set data into the artificial intelligence model for training and learning;

[0141] Artificial intelligence models are based on deep learning algorithms or machine learning algorithms to learn patterns and features in the training set data;

[0142] By learning patterns and features from the training set data, the AI ​​model can predict the performance of the mobile phone motherboard.

[0143] The technical effect of the above content is: the integrated training set data is substituted into the artificial intelligence model for training and learning, and then training is performed based on a deep learning algorithm or a machine learning algorithm. The deep learning algorithm or the machine learning algorithm can process and analyze a large amount of data, thereby training an artificial intelligence model that can predict performance data. By processing and learning these data, the patterns and features in the training set data are learned, and the laws and patterns in the data are gradually identified, thereby enhancing the accuracy and generalization ability of the artificial intelligence model. By learning the patterns and features in the training set data, the artificial intelligence model can predict the performance of the mobile phone motherboard.

[0144] Step three specifically includes the following steps:

[0145] Performance evaluation: Perform performance evaluation on the AI ​​model after training, including the accuracy, stability, scalability and reliability of the AI ​​model. The evaluation results will be used to optimize and improve the AI ​​model;

[0146] Among them, the performance evaluation uses the cross-validation method to divide the training set data into subsets to evaluate the artificial intelligence model;

[0147] Cross-validation methods include:

[0148] Data set partitioning: Divide the training set data into multiple non-overlapping subsets;

[0149] Model evaluation: Based on the cross-validation method, multiple non-overlapping subsets of data are substituted into the artificial intelligence model, and the artificial intelligence model is repeatedly evaluated using different subsets of data to finally obtain the evaluation results;

[0150] Result analysis: Analyze the evaluation results to find out the factors that affect the performance of the AI ​​model, and then optimize and improve the AI ​​model;

[0151] Model optimization and improvement: Based on the evaluation results of the artificial intelligence model, the artificial intelligence model is optimized and improved, including improving the training method and adding or replacing features. The optimized artificial intelligence model will be used in the calibration of the mobile phone motherboard performance parameters in step 4.

[0152] The technical effect of the above content is: after the training of the artificial intelligence model is completed and before it is applied, the artificial intelligence model needs to be evaluated and optimized. The performance evaluation of the artificial intelligence model adopts the cross-validation method to divide the training set data into multiple non-overlapping subset data, and substitute the divided subset data into the artificial intelligence model. The artificial intelligence model is repeatedly evaluated through different subset data to finally obtain the evaluation results. Then, the evaluation results are analyzed to find out the factors affecting the performance of the artificial intelligence model, and then the artificial intelligence model is optimized and improved. The optimization and improvement can adopt methods such as improving training methods and adding or replacing features. By evaluating and optimizing the artificial intelligence model, the performance of the artificial intelligence model can be effectively evaluated and optimized to ensure that the artificial intelligence model can achieve the expected results in practical applications.

[0153] Step 4 specifically includes the following steps:

[0154] Apply the optimized AI model to calibrate the performance parameters of mobile phone motherboards;

[0155] By inputting the status data of the current mobile phone motherboard, the artificial intelligence model predicts the performance of the mobile phone motherboard;

[0156] Based on the prediction results, point out the optimization suggestions for the mobile phone motherboard;

[0157] Calibrate the performance parameters of the mobile phone motherboard according to the existing optimization suggestions and determine the parameters that need to be adjusted.

[0158] Step six specifically includes the following steps:

[0159] After the mobile phone mainboard has been parameter adjusted in step 5, actual data of the mobile phone mainboard is obtained;

[0160] Integrate the acquired actual data into test set data;

[0161] Substitute the integrated test set data into the artificial intelligence model for verification;

[0162] The test results are used to determine how well the AI ​​model’s predictions match the actual data.

[0163] The technical effect of the above content is: after the artificial intelligence model is optimized and improved, the optimized artificial intelligence model can be applied to the calibration of the performance parameters of the mobile phone motherboard. During calibration, it is necessary to obtain the status data of the current mobile phone motherboard and input it, so that the artificial intelligence model can predict the performance of the mobile phone motherboard, and then point out the optimization suggestions for the mobile phone motherboard based on the prediction results, calibrate the performance parameters of the mobile phone motherboard based on the existing optimization suggestions, and determine the parameters that need to be adjusted. By calibrating the performance parameters of the mobile phone motherboard, it can help predict the difference between the performance parameters of the mobile phone motherboard and the actual performance, and then complete the calibration of the performance parameters. After determining the difference between the performance parameters of the mobile phone motherboard and the actual performance, adjust the performance parameters according to the calibration results to ensure that the mobile phone motherboard meets the preset performance standards. After the parameters of the mobile phone motherboard are adjusted, the actual data of the mobile phone motherboard can be obtained and integrated into test set data. Then, the test set data is substituted into the artificial intelligence model for verification, and the degree of consistency between the prediction of the artificial intelligence model and the actual data is judged according to the test results. The effect of the calibration of the performance parameters of the mobile phone motherboard can be verified through testing, thereby ensuring the accuracy of the calibration of the performance parameters of the mobile phone motherboard.

[0164] Working principle: by collecting relevant performance data of mobile phone motherboards, pre-processing the collected performance data and integrating them into training set data, at the same time, establishing an artificial intelligence model, substituting the training set data into the artificial intelligence model for training and learning, by training the artificial intelligence model, and evaluating and optimizing the trained artificial intelligence model, the accuracy of the artificial intelligence model prediction can be improved, and applying the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone motherboard can help the difference between the performance parameters of the mobile phone motherboard predicted by the artificial intelligence model and the actual performance. By evaluating and optimizing the artificial intelligence model, the performance of the artificial intelligence model can be effectively evaluated and optimized to ensure that the artificial intelligence model can achieve the expected effect in practical applications. The optimized artificial intelligence model is applied to the calibration of the performance parameters of the mobile phone motherboard to realize the calibration of the performance parameters of the mobile phone motherboard. Finally, the mobile phone motherboard with adjusted parameters is tested, and the degree of consistency between the prediction of the artificial intelligence model and the actual data is judged according to the test results, thereby ensuring the accuracy of the calibration of the performance parameters of the mobile phone motherboard.

[0165] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0166] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence, characterized in that: The following steps are involved: Step 1: Collect relevant performance data of mobile phone motherboards, and at the same time, establish an artificial intelligence model to pre-process the collected performance data of mobile phone motherboards and integrate them into training set data; Step 2: Use the training set data to train the artificial intelligence model, so that the trained artificial intelligence model can predict the performance of the mobile phone motherboard; Step 3: After training is completed, the AI ​​model is evaluated and optimized based on the evaluation results; Step 4: Apply the optimized artificial intelligence model to the calibration of the performance parameters of the mobile phone motherboard to achieve the calibration of the performance parameters of the mobile phone motherboard; Step 5: Based on the prediction of the artificial intelligence model, adjust the parameters of the mobile phone motherboard to optimize the performance of the mobile phone motherboard; Step 6: Test the mobile phone motherboard after parameter adjustment to verify the effect of the mobile phone motherboard performance parameter calibration through testing.

2. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 1, characterized in that: The step 1 specifically includes the following steps: Data collection: collect relevant performance data of the mobile phone motherboard through special sensors or test tools. The collected performance data is the basis of model training; Model building: Building artificial intelligence models based on the collected performance data; Data preprocessing: preprocess the collected performance data, including data cleaning, data conversion and data reduction; Data integration: The pre-processed performance data is integrated into training set data, which is used to train the established artificial intelligence model.

3. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 2, characterized in that: The preprocessing comprises: Data cleaning: remove errors, missing values, and noise from performance data; Data conversion: converting performance data into a unified data format or data type; Data reduction: Reduce the dimensionality of performance data through feature extraction, reduce computational costs and improve AI model performance.

4. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 3, characterized in that: The preprocessing further comprises: Extract the data information before removing the noise; According to the value of the data information before the noise is removed, obtaining a data distribution diagram of the data information before the noise is removed as a first data distribution diagram; According to the value of the data information after the noise is removed, a data distribution diagram of the data information after the noise is removed is obtained as a second data distribution diagram; Compare the first data distribution graph and the second data distribution graph to obtain a similarity value between the first data distribution graph and the second data distribution graph; Comparing the similarity value between the first data distribution graph and the second data distribution graph with a preset similarity threshold; When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, performing data quality evaluation on the data information after noise removal; When the similarity value between the first data distribution graph and the second data distribution graph is lower than a preset similarity threshold, traversing each data information of the first data distribution graph and each data information of the second data distribution graph, and obtaining the distribution position of each data information in the first data distribution graph and the second data distribution graph; Obtaining a distribution position straight-line distance according to the distribution position of each piece of data information in the first data distribution graph and the second data distribution graph; The data distortion coefficient is obtained according to the straight-line distance corresponding to the distribution position of each data information in the first data distribution diagram and the second data distribution diagram; wherein the data distortion coefficient is obtained by the following formula: Wherein, S represents the data distortion coefficient; k represents the number of data information contained in the first data distribution diagram and the second data distribution diagram; L i Indicates the straight-line distance of the distribution location corresponding to the i-th data information; L yi represents the minimum straight-line distance allowed for the distribution position corresponding to the ith data information that meets the similarity threshold; m represents the number of data information whose straight-line distances of the distribution positions in the k data information exceed the minimum straight-line distance allowed for the corresponding distribution positions; n represents the number of data information whose straight-line distances of the distribution positions in the k data information are 0; q represents the number of data information whose straight-line distances of the distribution positions in the k data information do not exceed the minimum straight-line distance allowed for the corresponding distribution positions, and whose straight-line distances are non-zero; s 01 、s 02 and 03 Respectively represent the first adjustment coefficient, the second adjustment coefficient and the third adjustment coefficient; The first adjustment coefficient is obtained by the following formula: Among them, s 01 represents the first adjustment coefficient; m represents the number of data information whose straight-line distance of the distribution position among the k data information exceeds the minimum straight-line distance allowed by the corresponding distribution position; L 01i Indicates the straight-line distance corresponding to the data information whose straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed by its corresponding distribution position; L y01i The minimum straight-line distance allowed for the distribution position corresponding to the data information indicating that the straight-line distance of the i-th distribution position exceeds the minimum straight-line distance allowed for the corresponding distribution position; At the same time, the second adjustment coefficient is obtained by the following formula: Among them, s 02 represents the second adjustment coefficient; n represents the number of data information whose straight-line distance of distribution position is 0 among the k data information; L 02xi and L 02hi They respectively represent the straight-line distance corresponding to the previous data information and the straight-line distance corresponding to the next data information adjacent to the data information whose straight-line distance to the i-th distribution position is 0; L yx02i and L yh02i They respectively represent the minimum straight-line distance allowed at the distribution position corresponding to the previous data information and the minimum straight-line distance allowed at the distribution position corresponding to the next data information; At the same time, the third adjustment coefficient is obtained by the following formula: Among them, s 03 represents the third adjustment coefficient; q represents the number of data information whose straight-line distance of the distribution position in the k data information does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and whose straight-line distance is non-zero; L 03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by its corresponding distribution position, and the straight-line distance corresponds to the data information with non-zero straight-line distance; L y03i Indicates that the straight-line distance of the ith distribution position does not exceed the minimum straight-line distance allowed by the corresponding distribution position, and the minimum straight-line distance allowed by the distribution position corresponding to the data information with non-zero straight-line distance; When the data distortion coefficient exceeds a preset coefficient threshold, an abnormal denoising operation alarm is issued; When the data distortion coefficient does not exceed a preset coefficient threshold, the data information before noise removal is re-cleaned.

5. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 4, characterized in that: When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, a data quality evaluation is performed on the data information after the noise is removed, including: When the similarity value between the first data distribution graph and the second data distribution graph is not less than a preset similarity threshold, extracting the signal-to-noise ratio, data integrity and noise residue of the data information after the noise is removed; The data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by using the signal-to-noise ratio, data integrity and noise residual of the data information after the noise is removed; wherein the data quality evaluation parameter corresponding to the data information after the noise is removed is obtained by the following formula: Where W represents the data quality evaluation parameter corresponding to the data information after noise removal; p represents the data integrity of the i-th data information; R i represents the noise residual corresponding to the i-th data information; C i represents the maximum allowable noise residual corresponding to the i-th data information; B i represents the signal-to-noise ratio of the data information after noise removal corresponding to the i-th data information; B xi represents the signal-to-noise ratio of the data information before noise removal corresponding to the i-th data information; Comparing the data quality evaluation parameter corresponding to the data information after noise removal with a preset data quality evaluation parameter threshold; When the data quality evaluation parameter corresponding to the data information after noise removal is lower than a preset data quality evaluation parameter threshold, a data quality abnormality alarm is issued.

6. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 1, characterized in that: The step 2 specifically includes the following steps: Substitute the integrated training set data into the artificial intelligence model for training and learning; Artificial intelligence models are based on deep learning algorithms or machine learning algorithms to learn patterns and features in the training set data; By learning patterns and features from the training set data, the AI ​​model can predict the performance of the mobile phone motherboard.

7. The method for calibrating mobile phone motherboard performance parameters based on artificial intelligence according to claim 1, characterized in that: The step three specifically includes the following steps: Performance evaluation: Perform performance evaluation on the AI ​​model after training. The evaluation results will be used to optimize and improve the AI ​​model; Among them, the performance evaluation uses the cross-validation method to divide the training set data into subsets to evaluate the artificial intelligence model; Among them, the cross-validation method includes: Data set partitioning: Divide the training set data into multiple non-overlapping subsets; Model evaluation: Based on the cross-validation method, multiple non-overlapping subsets of data are substituted into the artificial intelligence model, and the artificial intelligence model is repeatedly evaluated using different subsets of data to finally obtain the evaluation results; Result analysis: Analyze the evaluation results to find out the factors that affect the performance of the AI ​​model, and then optimize and improve the AI ​​model; Model optimization and improvement: Based on the evaluation results of the artificial intelligence model, the artificial intelligence model is optimized and improved. The optimized artificial intelligence model will be used in the calibration of the performance parameters of the mobile phone motherboard in step four.

8. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 7, characterized in that: Performance evaluation of AI models, including evaluation of the accuracy, stability, scalability and reliability of AI models; optimization and improvement of AI models, including improving training methods and adding or replacing features.

9. The method for calibrating performance parameters of a mobile phone motherboard based on artificial intelligence according to claim 1, characterized in that: The step 4 specifically includes the following steps: Apply the optimized AI model to calibrate the performance parameters of mobile phone motherboards; By inputting the status data of the current mobile phone motherboard, the artificial intelligence model predicts the performance of the mobile phone motherboard; Based on the prediction results, point out the optimization suggestions for the mobile phone motherboard; Calibrate the performance parameters of the mobile phone motherboard according to the existing optimization suggestions and determine the parameters that need to be adjusted.

10. The method for calibrating mobile phone motherboard performance parameters based on artificial intelligence according to claim 1, characterized in that: The step six specifically includes the following steps: After the mobile phone mainboard has been parameter adjusted in step 5, actual data of the mobile phone mainboard is obtained; Integrate the acquired actual data into test set data; Substitute the integrated test set data into the artificial intelligence model for verification; The test results are used to determine how well the AI ​​model’s predictions match the actual data.

Images