Scoring method and terminal for bimodal distribution vehicle detection indexes

Abstract

The invention discloses a scoring method for bimodal distribution vehicle detection indexes and a terminal. The scoring method comprises the following steps: acquiring bimodal distribution vehicle detection index data and vehicle information data; grouping the vehicle detection index data according to a classification criterion selected from the vehicle information data and acquiring grouped data; clustering the grouped data by using a k-means algorithm, setting the category to be 2, and obtaining clustered category grouped data; performing distribution fitting calculation on grouped data of each clustering category by using t distribution; obtaining a probability density function of the grouped data of each clustering category according to a distribution fitting calculation result, and recording a vehicle detection index data value corresponding to the highest point in the probability density function as a mark value; each type of vehicle detection index data is analyzed, and the vehicle detection index data is converted into a vehicle score by taking a flag value as a full score point in combination with a cumulative or residual distribution function.

Description

Technical Field

[0001] The present invention relates to the field of vehicle performance evaluation, and in particular to a scoring method and a terminal for vehicle detection indicators with bimodal distribution. Background Art

[0002] Due to the continuous reduction of traditional energy and environmental pollution, the utilization and development of new energy has been raised to a new level. The popularity of electric vehicles is getting higher and higher. Batteries are the core components of electric vehicles. The quantitative scoring of vehicle inspection indicators is an important part of the inspection report. A reasonable scoring scheme is conducive to users, car dealers and users to intuitively evaluate the value and quality of the vehicle. In vehicle performance evaluation and testing, it is necessary to quantify and horizontally compare the performance indicators of different vehicles, but the existing technology has the following problems:

[0003] (1) Unable to accurately identify the characteristics of bimodal distribution data: When faced with bimodal distribution data, traditional methods find it difficult to effectively identify and process the bimodal characteristics of the data, resulting in large deviations in the scoring results.

[0004] (2) The scoring method lacks scientificity and rationality: When dealing with bimodal distribution data, existing scoring methods often use simple linear transformation or single distribution model, which fails to fully consider the actual distribution characteristics of the data. The scoring results are difficult to reflect the true situation of the data. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a scoring method and terminal for vehicle detection indicators with a bimodal distribution, so as to solve the problem that vehicle detection indicator data with a bimodal distribution cannot be quantified for horizontal comparison.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A scoring method for bimodal distribution vehicle detection indicators, characterized in that it includes the steps of:

[0008] S1. Obtain bimodal distributed vehicle detection index data and vehicle information data;

[0009] S2, grouping the vehicle detection index data according to a classification standard selected from the vehicle information data and obtaining grouped data;

[0010] S3, clustering the grouped data using a k-means algorithm, with the category set to 2, to obtain cluster category grouped data;

[0011] S4, using t distribution to perform distribution fitting calculation on each of the cluster category grouping data;

[0012] S5. Obtain a probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as a mark value;

[0013] S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

[0014] In order to solve the above technical problems, another technical solution adopted by the present invention is:

[0015] A scoring terminal for bimodal distribution vehicle detection indicators, characterized by: comprising a client;

[0016] The invention comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are performed:

[0017] S1. Obtain bimodal distributed vehicle detection index data and vehicle information data;

[0018] S2, grouping the vehicle detection index data according to a classification standard selected from the vehicle information data and obtaining grouped data;

[0019] S3, clustering the grouped data using a k-means algorithm, with the category set to 2, to obtain cluster category grouped data;

[0020] S4, using t distribution to perform distribution fitting calculation on each of the cluster category grouping data;

[0021] S5. Obtain a probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as a mark value;

[0022] S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

[0023] The beneficial effects of the present invention include at least: by adopting a scientific and reasonable clustering and scoring method for bimodal distribution of vehicle detection index data, the problem that traditional scoring methods are difficult to handle bimodal distribution data is solved, and the accuracy and objectivity of scoring are improved, as follows:

[0024] (1) The k-means clustering algorithm can effectively divide the bimodal distribution data into two cluster categories and accurately identify the two concentrated peaks of the data. In this way, the bimodal characteristics of the data can be accurately captured, improving the accuracy of data analysis.

[0025] (2) By fitting the t distribution to the data of each cluster category, the probability density function is obtained, and the mark value is determined as the highest density point, so that the scoring method is scientific and reasonable. The cumulative distribution function (CDF) and residual distribution function (SF) are used for scoring conversion to ensure that the scoring method can reflect the distribution characteristics and actual situation of the data, making the scoring results more objective and fair.

[0026] (3) By converting vehicle inspection index data into a unified score, the performance of different vehicles can be quantified and compared horizontally, which improves the comparability and analytical value of the data. This helps manufacturers, regulators and users to more intuitively evaluate and compare vehicle performance, and provides a scientific basis for optimizing vehicle design and improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a scoring method for a bimodal distribution vehicle detection index in an embodiment of the present invention;

[0028] Figure 2 A specific flow chart of a scoring method for a bimodal distribution vehicle detection index in an embodiment of the present invention;

[0029] Figure 3 is a schematic diagram of a probability density function in an embodiment of the present invention;

[0030] Figure 4 is a schematic diagram of a cumulative distribution function in an embodiment of the present invention;

[0031] Figure 5 is a schematic diagram of a residual distribution function in an embodiment of the present invention;

[0032] Figure 6 A schematic diagram of a scoring terminal for a vehicle detection index with bimodal distribution in an embodiment of the present invention;

[0033] Description of labels:

[0034] 1. A scoring terminal for vehicle detection indicators with bimodal distribution; 2. A memory; 3. A processor. DETAILED DESCRIPTION

[0035] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.

[0036] The relevant terms in the present invention are explained as follows:

[0037] VIN code: Vehicle identification number.

[0038] t distribution: t distribution is a probability distribution in statistics, often used to infer population parameters in small sample cases.

[0039] Probability density function: The probability density function (PDF) of the t distribution describes the probability distribution of a random variable that follows the t distribution.

[0040] Cumulative Distribution Function: The Cumulative Distribution Function (CDF) of the t-distribution describes the probability that a random variable is less than or equal to a given value.

[0041] Survival function (SF) of t distribution describes the probability that a random variable is greater than a given value. For t distribution, its SF is the complement of the cumulative distribution function.

[0042] k-means: k-means is a commonly used clustering algorithm that is used to divide data points into K clusters with similar characteristics. This algorithm is widely used in data mining, machine learning, and pattern recognition.

[0043] Please refer to Figures 1 to 5 , a scoring method for a bimodal distribution vehicle detection index, comprising the steps of:

[0044] S1. Obtain bimodal distributed vehicle detection index data and vehicle information data;

[0045] S2, grouping the vehicle detection index data according to a classification standard selected from the vehicle information data and obtaining grouped data;

[0046] S3, clustering the grouped data using a k-means algorithm, with the category set to 2, to obtain cluster category grouped data;

[0047] S4, using t distribution to perform distribution fitting calculation on each of the cluster category grouping data;

[0048] S5. Obtain a probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as a mark value;

[0049] S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

[0050] From the above description, it can be seen that the beneficial effects of the present invention are: by adopting a scientific and reasonable clustering and scoring method for bimodal distribution of vehicle detection index data, the problem that traditional scoring methods are difficult to handle bimodal distribution data is solved, and the accuracy and objectivity of scoring are improved, as follows:

[0051] (1) The k-means clustering algorithm can effectively divide the bimodal distribution data into two cluster categories and accurately identify the two concentrated peaks of the data. In this way, the bimodal characteristics of the data can be accurately captured, improving the accuracy of data analysis.

[0052] (2) By fitting the t distribution to the data of each cluster category, the probability density function is obtained, and the mark value is determined as the highest density point, so that the scoring method is scientific and reasonable. The cumulative distribution function (CDF) and residual distribution function (SF) are used for scoring conversion to ensure that the scoring method can reflect the distribution characteristics and actual situation of the data, making the scoring results more objective and fair.

[0053] (3) By converting vehicle inspection index data into a unified score, the performance of different vehicles can be quantified and compared horizontally, which improves the comparability and analytical value of the data. This helps manufacturers, regulators and users to more intuitively evaluate and compare vehicle performance, and provides a scientific basis for optimizing vehicle design and improvement.

[0054] It can be understood that the k-means clustering algorithm is used in this example to divide the data points into 2 clusters with similar characteristics, as shown below:

[0055] The original data of a detection index is: [0.4677, -0.5033, -0.7014, 0.5333, -0.73, 0.9967, -0.5502, -0.0243, 0.9136, 0.7154, -0.7161, 0.4515, -0.7514, 0.2203, 0.6211, 0.6285, 0.3787, -0.5347, 0.3325, 0 .4577, -0.5236, 0.6816, -0.6419, -0.4345, -0.5078, -0.9138, 0.3545, -0.254, 0.4425, -0.5959, 0.6962, -0.6338, -0.7287, 0.7266, -0.2035, -0.3229, -0.5548, -0.4591, -0.5677, - 0.5242, 0.2678, -0.5256, 0.6846, 0.5859, -0.4398, 0.5823, -0.7559, 0.3177, -0.454, 0.5283, -0.4285, 0.6235, 0.5064, 0.7627, -0.2099, 0.6188, 0.4866, -0.4998, 0.9367, 0.6197 , -0.6477, 0.2365, 0.6203, 0.5773, 0.6114, 0.7466, 0.4806, 0.9441, -0.3759, -0.3283, 0.2831, -0.6406, -0.5484, -0.8568, -0.5285, -0.8205, -0.7086, -0.4961, 0.5227, -0.2896];

[0056] Using the k-means algorithm, the number of clusters is set to 2, and two cluster groups are obtained:

[0057] Cluster grouping 1: [-0.5033, -0.7014, -0.73, -0.5502, -0.0243, -0.7161, -0.7514, -0.5347, -0.5236, -0.6419, -0.4345, -0.5078, -0.9138, -0.254, -0.5959, -0.6338, -0.7287, -0.2035, -0.3229, -0.5548, - 0.4591, -0.5677, -0.5242, -0.5256, -0.4398, -0.7559, -0.454, -0.4285, -0.2099, -0.4998, -0.6477, -0.3759, -0.3283, -0.6406, -0.5484, -0.8568, -0.5285, -0.8205, -0.7086, -0.4961, -0.2896]

[0058] In cluster group 1, the cluster center calculated by the algorithm is: -0.53492927;

[0059] Cluster group 2: [0.4677, 0.5333, 0.9967, 0.9136, 0.7154, 0.4515, 0.2203, 0.6211, 0.6285, 0.3787, 0.3325, 0.4577, 0.6816, 0.3545, 0.4425, 0.6962, 0.7266, 0.2678, 0.6846, 0.5859, 0.5823, 0.3177, 0.5283, 0.6235, 0.5064, 0.7627, 0.6188, 0.4866, 0.9367, 0.6197, 0.2365, 0.6203, 0.5773, 0.6114, 0.7466, 0.4806, 0.9441, 0.2831, 0.5227]

[0060] In cluster grouping 2, the cluster center calculated by the algorithm is: 0.56825641.

[0061] Please refer to Figures 4 to 5 In an embodiment of the present invention, step S6 specifically includes the following steps:

[0062] S61, analyzing each of the vehicle detection index data, if the corresponding score shows an increasing trend when the vehicle detection index data value increases, then proceeding to step S62; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, then proceeding to step S63;

[0063] S62. Taking the mark value as the full score and combining it with the cumulative distribution function, a vehicle score calculation formula is obtained:

[0064] sCore=a+b*log(p).

[0065] In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score;

[0066] The vehicle detection index data is converted into a vehicle score using the vehicle score calculation formula;

[0067] S63, taking the mark value as the full score and combining it with the residual distribution function to obtain the vehicle score calculation formula:

[0068] sCore=a+b*log(p).

[0069] In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score;

[0070] The vehicle scoring calculation formula is used to convert the vehicle detection index data into a vehicle score.

[0071] From the above description, it can be seen that in step S6, by analyzing the correlation between the central value of the cluster category to which each vehicle detection index data belongs and the score, it is automatically determined whether the data is showing an increasing trend or a decreasing trend (negative growth trend), so as to select the corresponding score conversion method to ensure the accuracy of the score; the mark value is used as the full score point, and the score calculation formula of the cumulative distribution function or the residual distribution function is used, which is applicable to various vehicle detection index data, and effective score conversion can be performed regardless of their distribution characteristics.

[0072] In an embodiment of the present invention, the step S62 further includes the steps of:

[0073] The input vehicle detection index data is judged. If the vehicle detection index data is greater than the mark value, the vehicle score is deemed to be full marks; if the vehicle detection index data is less than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

[0074] From the above description, it can be seen that in step S5, the vehicle detection index data value corresponding to the highest point in the probability density function is recorded as the mark value, that is, after the cluster category grouping data is subjected to distribution fitting calculation, the highest point extracted from the calculation result represents the optimal value of this type of detection index value, and the optimal value is the full score point as the scoring limit. Under the above-mentioned positive correlation condition, when the vehicle detection index data is greater than the mark value, it is also regarded as a full score. When the vehicle detection index data is less than the mark value, it is substituted into the vehicle scoring calculation formula to calculate the actual score; however, when a negative value appears according to the vehicle scoring calculation formula, the score is regarded as zero for the convenience of display and statistics.

[0075] In an embodiment of the present invention, the step S63 further includes the steps of:

[0076] The input vehicle detection index data is judged. If the vehicle detection index data is less than the mark value, the vehicle score is deemed to be full marks; if the vehicle detection index data is greater than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

[0077] From the above description, it can be seen that in step S5, the vehicle detection index data value corresponding to the highest point in the probability density function is recorded as the mark value, that is, after the cluster category grouping data is calculated through distribution fitting, the highest point extracted from the calculation result represents the optimal value of this type of detection index value, and the optimal value is the full score point as the scoring limit. Under the above-mentioned negative correlation condition, when the vehicle detection index data is less than the mark value, it is also regarded as a full score. When the vehicle detection index data is greater than the mark value, it is substituted into the vehicle scoring calculation formula to calculate the actual score; however, when a negative value appears according to the vehicle scoring calculation formula, the score is regarded as zero for the convenience of display and statistics.

[0078] That is, when the test index data is greater than or less than the marked value, the full score and zero score are set reasonably to ensure the rationality and consistency of the scoring results.

[0079] In an embodiment of the present invention, the step S1 further comprises the steps of:

[0080] The vehicle information data is extracted using the vehicle VIN code.

[0081] From the above description, it can be seen that extracting vehicle information data through the vehicle VIN code can accurately associate the detailed information of each vehicle and provide a reliable data basis for subsequent grouping and clustering.

[0082] Please refer to Figure 6A scoring terminal 1 for a vehicle detection index with a bimodal distribution includes a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, the steps in a scoring method for a vehicle detection index with a bimodal distribution are completed.

[0083] The present invention provides a scoring method and terminal for vehicle detection indexes with bimodal distribution, which are mainly used to quantify vehicle detection index data with bimodal distribution and enable horizontal comparison between different vehicles.

[0084] Please refer to Figures 1 to 5 , Embodiment 1 of the present invention is:

[0085] A scoring method for a bimodal distribution vehicle detection index comprises the steps of:

[0086] S1. Obtain bimodal distributed vehicle detection index data and vehicle information data;

[0087] S2, grouping the vehicle detection index data according to the classification standard selected from the vehicle information data and obtaining grouped data;

[0088] S3, clustering the grouped data using the k-means algorithm, with the category set to 2, to obtain cluster category grouped data;

[0089] S4, using t distribution to perform distribution fitting calculation on each cluster category grouping data;

[0090] S5. Obtain the probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as the mark value;

[0091] S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

[0092] Specifically, in step S1, the vehicle information data is extracted using the vehicle VIN code.

[0093] That is, in this embodiment, by using a scientific and reasonable clustering and scoring method for the bimodal distribution of vehicle detection index data, the problem that the traditional scoring method is difficult to handle bimodal distribution data is solved, and the accuracy and objectivity of the scoring are improved, as follows:

[0094] (1) The k-means clustering algorithm can effectively divide the bimodal distribution data into two cluster categories and accurately identify the two concentrated peaks of the data. In this way, the bimodal characteristics of the data can be accurately captured, improving the accuracy of data analysis.

[0095] (2) By fitting the t distribution to the data of each cluster category, the probability density function is obtained, and the mark value is determined as the highest density point, so that the scoring method is scientific and reasonable. The cumulative distribution function (CDF) and residual distribution function (SF) are used for scoring conversion to ensure that the scoring method can reflect the distribution characteristics and actual situation of the data, making the scoring results more objective and fair.

[0096] (3) By converting vehicle inspection index data into a unified score, the performance of different vehicles can be quantified and compared horizontally, which improves the comparability and analytical value of the data. This helps manufacturers, regulators and users to more intuitively evaluate and compare vehicle performance, and provides a scientific basis for optimizing vehicle design and improvement.

[0097] Please refer to Figures 4 to 5 , Embodiment 2 of the present invention is:

[0098] Based on the first embodiment, step S6 specifically includes the following steps:

[0099] S61, analyzing each vehicle detection index data, if the corresponding score shows an increasing trend when the vehicle detection index data value increases, then proceed to step S62; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, then proceed to step S63;

[0100] S62. Taking the mark value as the full score, combined with the cumulative distribution function, the vehicle score calculation formula is obtained:

[0101] sCore=a+b*log(p).

[0102] In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated based on the relationship between the mark value and the full score;

[0103] The vehicle scoring calculation formula is used to convert the vehicle inspection index data into a vehicle score;

[0104] Among them, the input vehicle inspection index data is judged. If the vehicle inspection index data is greater than the mark value, the vehicle score is regarded as full marks; if the vehicle inspection index data is less than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is regarded as zero.

[0105] S63. Taking the mark value as the full score and combining it with the residual distribution function, the vehicle score calculation formula is obtained:

[0106] score=a+b*log(p).

[0107] In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated based on the relationship between the mark value and the full score;

[0108] The vehicle scoring calculation formula is used to convert vehicle inspection index data into vehicle scores.

[0109] Among them, the input vehicle inspection index data is judged. If the vehicle inspection index data is less than the mark value, the vehicle score is regarded as full marks; if the vehicle inspection index data is greater than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is regarded as zero.

[0110] Specifically, the calculation examples of parameters a and b are as follows:

[0111] Define the mark value p as the full score, and the full score is 100 points; select the median point according to the nature of the vehicle detection index data and the vehicle information data, for example, define the score at 0.1 times p as 50 points, and bring the two sets of data into the vehicle score calculation formula to obtain the specific values ​​of a and b, and then get the corresponding vehicle score based on the actual input vehicle detection index data.

[0112] Please refer to Figure 2 , Embodiment 3 of the present invention is:

[0113] An example of application of a scoring method for bimodal distribution vehicle detection indicators is as follows:

[0114] S1. Extract corresponding vehicle inspection index data from the vehicle inspection order, and extract vehicle information data such as vehicle battery type, manufacturer, model, brand data, etc. according to the vehicle VIN code;

[0115] S2. Grouping the index data according to the specified vehicle information classification field, where the classification field is selected from the vehicle information. In this example, the vehicle model is used as the classification field;

[0116] The process is as follows:

[0117] Get vehicle information and test index data based on VIN code:

[0118]

[0119]

[0120] Now, select "Model" as the classification field to group the inspection metric data for these vehicles by model:

[0121] Grouping results:

[0122] Vehicle Type: SUV

[0123] VIN: 1A2B3C4D, detection index (voltage accuracy): 8.5

[0124] VIN: 3C4D5E6F, detection index (voltage accuracy): 9.0

[0125] VIN: 5E6F7G8H, detection index (voltage accuracy): 7.8

[0126] Model: Sedan

[0127] VIN: 2B3C4D5E, detection index (voltage accuracy): 6.2

[0128] VIN: 4D5E6F7G, detection index (voltage accuracy): 6.5

[0129] VIN: 6F7G8H9I, detection index (voltage accuracy): 6.0

[0130] S3. Cluster the grouping index data based on the k-means algorithm, set the category to 2, and the data in each cluster category belongs to the data of one peak in the bimodal category, so that the bimodal characteristic data is converted into single-peak characteristic data under two cluster categories;

[0131] S4, grouping the data according to each cluster category, and using t distribution to perform distribution fitting calculation on the detection index;

[0132] S5. Obtain the probability density function of each cluster category of the detection index from the distribution fitting calculation results. The detection index value corresponding to the point with the highest density is recorded as peak_loc (mark value), such as Figure 3 ;

[0133] S6, please refer to Figure 4 , if the detection index value belongs to a small classification cluster center value, and in the cluster data, the larger the detection index value, the better, it is converted into a score according to the following process:

[0134] 1) Use the cumulative distribution function of the detection index for calculation, such as Figure 4 As shown;

[0135] 2) According to the peak_loc obtained in step 4, obtain the corresponding cumulative probability, denoted as p;

[0136] 3) Define the score at probability p as 100 points and the score at 0.1p as 50 points, according to the formula:

[0137] score = a + b * log (p);

[0138] Substitute the values ​​of coefficients a and b into the calculation to obtain the calculation formula for converting probability p to score. Then, the detection index value can be converted into cumulative probability first and then into score to complete the conversion of detection index value to score.

[0139] 4) If the detection index value is greater than peak_loc, the score is fixed to the highest score of 100 points; if it is less than or equal to peak_loc, the score calculation formula is used to convert the cumulative probability corresponding to the detection index value into a calculated score. If score<0, the score is fixed to the lowest score of 0 points, otherwise the calculated converted score is taken as the final conversion score.

[0140] Please refer to Figure 5 , if the detection index value belongs to a large classification cluster center value, and in the cluster data, the smaller the detection index value, the better, it is converted into a score according to the following process:

[0141] 1) Calculate using the residual distribution function of the detection index;

[0142] 2) According to the peak_loc obtained in step 4, obtain the corresponding residual probability, denoted as p;

[0143] 3) Define the score at probability p as 100 points and the score at 0.1p as 50 points, according to the formula:

[0144] score = a + b * log (p);

[0145] Substitute the values ​​of coefficients a and b into the calculation to obtain the calculation formula for converting probability p to score. Then, the detection index value can be converted into cumulative probability first and then into score to complete the conversion of detection index value to score.

[0146] 4) If the detection index value is less than peak_loc, the score is fixed to the highest score of 100 points; if it is greater than or equal to peak_loc, the score calculation formula is used to convert the cumulative probability corresponding to the detection index value into a calculated score. If score<0, the score is fixed to the lowest score of 0 points, otherwise the calculated converted score is taken as the final conversion score.

[0147] Preferably, the method further comprises step S6:

[0148] Get the scores of each indicator, and get the comprehensive score by summing up the indicator weights, as follows:

[0149] score 0 =∑w factor *score factor ;

[0150] Among them, score 0 is the comprehensive score, w factor is the indicator scoring weight, score factor is the index score, i.e. the score obtained by the above vehicle score calculation formula.

[0151] Please refer to Figure 6 , Embodiment 4 of the present invention is: a scoring terminal 1 for a vehicle detection index with a bimodal distribution, comprising a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, the steps in the scoring method for a vehicle detection index with a bimodal distribution in any one of Embodiments 1 to 3 are completed.

[0152] In summary, the present invention provides a scoring method and terminal for vehicle detection indicators with bimodal distribution, which solves the problem that traditional scoring methods are difficult to handle bimodal distribution data by adopting scientific and reasonable clustering and scoring methods for vehicle detection indicator data with bimodal distribution, and improves the accuracy and objectivity of scoring, as follows:

[0153] (1) The k-means clustering algorithm can effectively divide the bimodal distribution data into two cluster categories and accurately identify the two concentrated peaks of the data. In this way, the bimodal characteristics of the data can be accurately captured, improving the accuracy of data analysis.

[0154] (2) By fitting the t distribution to the data of each cluster category, the probability density function is obtained, and the mark value is determined as the highest density point, so that the scoring method is scientific and reasonable. The cumulative distribution function (CDF) and residual distribution function (SF) are used for scoring conversion to ensure that the scoring method can reflect the distribution characteristics and actual situation of the data, making the scoring results more objective and fair.

[0155] (3) By converting vehicle inspection index data into a unified score, the performance of different vehicles can be quantified and compared horizontally, which improves the comparability and analytical value of the data. This helps manufacturers, regulators and users to more intuitively evaluate and compare vehicle performance, and provides a scientific basis for optimizing vehicle design and improvement.

[0156] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A scoring method for bimodal distribution vehicle detection indicators, characterized in that: Includes steps: S1. Obtain bimodal distributed vehicle detection index data and vehicle information data; S2, grouping the vehicle detection index data according to a classification standard selected from the vehicle information data and obtaining grouped data; S3, clustering the grouped data using a k-means algorithm, with the category set to 2, to obtain cluster category grouped data; S4, using t distribution to perform distribution fitting calculation on each of the cluster category grouping data; S5. Obtain a probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as a mark value; S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

2. The scoring method for a bimodal distribution vehicle detection index according to claim 1 is characterized in that: The step S6 specifically comprises the steps of: S61, analyzing each of the vehicle detection index data, if the corresponding score shows an increasing trend when the vehicle detection index data value increases, then proceeding to step S62; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, then proceeding to step S63; S62. Taking the mark value as the full score and combining it with the cumulative distribution function, a vehicle score calculation formula is obtained: score = a + b * log (p); In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score; The vehicle detection index data is converted into a vehicle score using the vehicle score calculation formula; S63, taking the mark value as the full score and combining it with the residual distribution function to obtain the vehicle score calculation formula: score = a + b * log (p); In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score; The vehicle scoring calculation formula is used to convert the vehicle detection index data into a vehicle score.

3. The scoring method for a bimodal distribution vehicle detection index according to claim 2 is characterized in that: The step S62 further comprises the steps of: The input vehicle detection index data is judged, and if the vehicle detection index data is greater than the mark value, the vehicle score is deemed to be full score; If the vehicle detection index data is less than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

4. The scoring method for a bimodal distribution vehicle detection index according to claim 2 is characterized in that: The step S63 further comprises the steps of: The input vehicle detection index data is judged, and if the vehicle detection index data is less than the mark value, the vehicle score is deemed to be full score; If the vehicle detection index data is greater than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

5. The scoring method for a bimodal distribution vehicle detection index according to claim 1 is characterized in that: The step S1 further comprises the steps of: The vehicle information data is extracted using the vehicle VIN code.

6. A scoring terminal for bimodal distribution of vehicle detection indicators, characterized in that: Including the client; The invention comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are performed: S1. Obtain bimodal distributed vehicle detection index data and vehicle information data; S2, grouping the vehicle detection index data according to a classification standard selected from the vehicle information data and obtaining grouped data; S3, clustering the grouped data using a k-means algorithm, with the category set to 2, to obtain cluster category grouped data; S4, using t distribution to perform distribution fitting calculation on each of the cluster category grouping data; S5. Obtain a probability density function of each cluster category grouping data according to the distribution fitting calculation result and record the vehicle detection index data value corresponding to the highest point in the probability density function as a mark value; S6. Analyze each type of vehicle detection index data. If the corresponding score shows an increasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the cumulative distribution function; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, the vehicle detection index data is converted into a vehicle score using the mark value as the full score and in combination with the residual distribution function.

7. The scoring terminal for bimodal distribution of vehicle detection indicators according to claim 6, characterized in that: The step S6 specifically comprises the steps of: S61, analyzing each of the vehicle detection index data, if the corresponding score shows an increasing trend when the vehicle detection index data value increases, then proceeding to step S62; if the corresponding score shows a decreasing trend when the vehicle detection index data value increases, then proceeding to step S63; S62. Taking the mark value as the full score and combining it with the cumulative distribution function, a vehicle score calculation formula is obtained: score = a + b * log (p); In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score; The vehicle detection index data is converted into a vehicle score using the vehicle score calculation formula; S63, taking the mark value as the full score and combining it with the residual distribution function to obtain the vehicle score calculation formula: score = a + b * log (p); In the formula, score represents the vehicle score; p represents the vehicle detection index data; a and b represent the parameters calculated according to the relationship between the mark value and the full score; The vehicle scoring calculation formula is used to convert the vehicle detection index data into a vehicle score.

8. The scoring terminal for bimodal distribution of vehicle detection indicators according to claim 7, characterized in that: The step S62 further comprises the steps of: The input vehicle detection index data is judged, and if the vehicle detection index data is greater than the mark value, the vehicle score is deemed to be full score; If the vehicle detection index data is less than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

9. The scoring terminal for bimodal distribution of vehicle detection indicators according to claim 7, characterized in that: The step S63 further comprises the steps of: The input vehicle detection index data is judged, and if the vehicle detection index data is less than the mark value, the vehicle score is deemed to be full score; If the vehicle detection index data is greater than or equal to the mark value, the vehicle score calculation formula is used to calculate the actual vehicle score. If the detected vehicle score is a negative value, the vehicle score is deemed to be zero.

10. The scoring terminal for bimodal distribution vehicle detection index according to claim 6, characterized in that: The step S1 further comprises the steps of: The vehicle information data is extracted using the vehicle VIN code.

Images