Abnormality analysis method based on big data

Abstract

The invention discloses an anomaly analysis method based on big data, and the method comprises the following steps: 1, guaranteeing the comprehensiveness of data through multi-source data collection; 2, carrying out preprocessing and feature extraction on the multi-source data; 3, establishing a prediction model and evaluating; 4, performing real-time prediction and comparison on the traffic data; 5, the system gives out early warning according to the prediction result, and the data collecting and preprocessing module is used for collecting and processing multi-source data; the prediction model establishment and evaluation module is used for establishing a traffic accident model; the real-time prediction and early warning module is used for performing traffic accident prediction and early warning according to the real-time data; the model selection and training set division module is used for selecting a suitable prediction model and dividing historical traffic data into a training set and a verification set for model training and evaluation; the system has the characteristics of accurate traffic accident analysis and timely early warning.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to an anomaly analysis method based on big data. Background Art

[0002] There are many challenges and problems in the current traffic safety field, such as frequent traffic accidents, traffic congestion, and inefficient traffic management. Traditional traffic safety measures are often passive responses, lacking timely warning and active intervention mechanisms. Therefore, the emergence of an abnormal analysis method based on big data provides a new way to solve these problems.

[0003] With the rapid development and application of big data technology, a large amount of traffic data has emerged in the transportation field, including historical traffic data, real-time traffic data and other related data. These data contain rich traffic characteristics and laws, which can be used for the prediction and warning of traffic accidents.

[0004] However, the current traffic accident prediction and warning technology has some shortcomings. First, the data collection and integration are imperfect, resulting in the limitation of data sources and the instability of data quality. Second, the limitations of feature extraction and selection methods make it impossible to fully mine the effective information in the data. Third, the traditional prediction model has certain limitations in accuracy and generalization ability, and cannot meet the needs of accurate prediction in complex traffic environments. Therefore, it is necessary to design an abnormal analysis method based on big data that is efficient, accurate, and timely. Summary of the invention

[0005] The purpose of the present invention is to provide an anomaly analysis method based on big data to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solution: an abnormality analysis method based on big data, the method comprising the following steps:

[0007] Step 1: Ensure the comprehensiveness of data through multi-source data collection;

[0008] Step 2: Preprocess and extract features of multi-source data;

[0009] Step 3: Establish prediction model and evaluation;

[0010] Step 4: Real-time prediction and comparison of traffic data;

[0011] Step 5: The system issues an early warning based on the prediction results.

[0012] According to the above technical solution, the step of ensuring the comprehensiveness of data through multi-source data collection includes:

[0013] Collect historical traffic data and real-time data;

[0014] Synthesize information from multiple data sources.

[0015] According to the above technical solution, the step of integrating information from multiple data sources includes:

[0016] Weather forecast data can be obtained through meteorological departments or meteorological data providers. These data include weather information related to traffic safety, such as rainfall, visibility, and temperature. Ensure that accurate weather forecast data is obtained so that it can be correlated with traffic accident risks for analysis. Finally, by cooperating with relevant departments or construction companies, road construction information can be obtained, including ongoing or planned road construction, temporary traffic restrictions, etc. Road construction is one of the potential risk factors for traffic accidents. Including it in the analysis scope can improve prediction accuracy. Through the above detailed steps, multi-source data such as historical traffic data, real-time data, social media data, weather forecast data, and road construction information can be collected to provide a more comprehensive information basis for potential risk analysis of traffic accidents.

[0017] According to the above technical solution, the steps of preprocessing and feature extraction of multi-source data include:

[0018] Perform data preprocessing to ensure data accuracy and consistency;

[0019] Extract features from historical traffic data and real-time data;

[0020] The extracted feature data is converted into numerical values.

[0021] According to the above technical solution, the step of converting the extracted feature data into numerical values ​​includes:

[0022] Different attributes of the data are extracted as features. When the location is used as a feature, the location is represented by geocoding or latitude and longitude, and converted into numerical features. The time of the traffic accident is used as a feature, and time information such as hours, dates, and days of the week can be extracted and converted into numerical features. Weather information can be used as a feature, including weather factors such as rainfall, visibility, and temperature, and converted into numerical features. Road types are used as features, such as highways, urban roads, etc., and road types can be encoded as numerical features. Traffic flow and speed can be used as features. They can be calculated based on historical traffic data and real-time data. Through data preprocessing and feature extraction, representative and information-rich features can be extracted from the original data to provide reliable input for subsequent analysis and modeling. These features will be used as input variables of the prediction model to help the system identify potential risk areas for traffic accidents.

[0023] According to the above technical solution, the steps of establishing a prediction model and evaluating include:

[0024] Use decision tree algorithm to establish traffic accident prediction model;

[0025] Use supervised learning methods with historical traffic data as input features;

[0026] Use evaluation metrics to validate the predictive power of the model.

[0027] According to the above technical solution, the step of using the decision tree algorithm to establish a traffic accident prediction model includes:

[0028] During the training process, historical traffic data is first used as the training set. Each training sample consists of features such as location, time, weather, road type, traffic flow, etc., and corresponds to a label indicating whether a traffic accident has occurred at the location. The goal is to use these features to build a decision tree model that can accurately predict traffic accidents based on the conditions of the input features. In the process of building the decision tree, the splitting criterion and algorithm are used to determine the splitting method of each node. The splitting criterion is used to calculate the information gain or Gini index value of each feature to measure the importance of the feature for classification. According to the results of the criterion calculation, the feature with the maximum information gain or the minimum Gini index is selected as the splitting feature of the current node.

[0029] According to the above technical solution, the step of real-time prediction and comparison of traffic data includes:

[0030] Input real-time data into the prediction model to make real-time traffic accident predictions;

[0031] Compare the predicted results with the actual data to determine the areas of traffic accident risk.

[0032] According to the above technical solution, the step of the system issuing an early warning according to the prediction result includes:

[0033] Once the prediction model determines that there is a high risk of traffic accidents at a certain location, the system will trigger the early warning mechanism. The early warning information will be generated and customized according to factors such as location, time, and risk level. For example, through the traffic navigation system, drivers can receive voice prompts or text warnings during navigation, informing them that they are about to enter an area with a risk of traffic accidents, and are advised to drive carefully or choose an alternative route. Mobile applications can send early warning information to drivers via push notifications so that they can understand and adjust their driving plans in a timely manner. Electronic billboards can display relevant warning information on the road to remind drivers to pay attention to safety and abide by traffic rules. Through traffic navigation systems, mobile applications, electronic billboards, etc., early warning information is promptly conveyed to drivers and traffic management departments to remind them to pay attention to traffic accident risk areas and take corresponding safety measures.

[0034] According to the above technical solution, the system includes:

[0035] Data collection and preprocessing module, used for multi-source data collection and processing;

[0036] Prediction model building and evaluation module, used to build traffic accident models;

[0037] The real-time prediction and warning module is used to predict traffic accidents and issue warnings based on real-time data.

[0038] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention analyzes traffic data by providing a data collection and preprocessing module, a prediction model establishment and evaluation module, and a real-time prediction and warning module, thereby realizing the prediction of traffic accidents. First, historical traffic data and real-time data are collected through multi-source data, and the location, time, weather and other factors of the traffic accident are comprehensively analyzed. Next, data preprocessing and feature extraction are performed to ensure the accuracy and consistency of the data, and useful features are extracted therefrom. Then, a traffic accident prediction model is established, and machine learning and other methods are used to train and optimize the model, and the prediction ability of the model is evaluated. In the real-time prediction and comparison stage, real-time data is input into the model, and real-time prediction of potential traffic accident risk areas is performed, and compared with actual data. Finally, warning information is issued according to the prediction results, and drivers and traffic management departments are reminded to take corresponding measures through traffic navigation systems, mobile phone applications, electronic billboards, etc. This method effectively analyzes traffic data, and the prediction is accurate and the warning is timely. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0040] Figure 1 A flowchart of an abnormality analysis method based on big data provided in Example 1 of the present invention;

[0041] Figure 2 A schematic diagram of the module composition of the big data-based anomaly analysis system provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0042] 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.

[0043] Embodiment 1: Figure 1 This is a flowchart of an abnormality analysis method based on big data provided in Embodiment 1 of the present invention. This embodiment can be applied to the scenario of traffic accident prediction. This method can be executed by an abnormality analysis system based on big data provided in this embodiment. Figure 1 As shown, the method specifically comprises the following steps:

[0044] Step 1: Ensure the comprehensiveness of data through multi-source data collection;

[0045] In the embodiment of the present invention, in order to achieve the accuracy and comprehensiveness of the traffic accident warning system, it is first necessary to collect multi-source data, including historical traffic data, real-time data, social media data, weather forecast data, road construction information and other data sources;

[0046] For example, first, by cooperating with the traffic management department, police station or other relevant agencies to collect traffic accident records in the past period of time, including relevant information such as location, time, weather, etc., and ensure the integrity and accuracy of the data, then use traffic monitoring cameras, traffic sensors and other equipment to collect real-time data. These real-time data include vehicle flow, speed, road conditions and other information. By cooperating with relevant traffic management departments, the reliability and timeliness of real-time data can be ensured; in addition, through social media platforms such as Xiaohongshu and Weibo, information related to traffic accidents can be collected. By using keyword search technology, content related to traffic accidents posted by users can be obtained, such as traffic congestion reports, descriptions of accident witnesses, etc.;

[0047] For example, weather forecast data can be obtained through meteorological departments or meteorological data providers. These data include weather information related to traffic safety, such as rainfall, visibility, and temperature. Accurate weather forecast data can be obtained so that it can be correlated with traffic accident risks for analysis. Finally, road construction information can be obtained by cooperating with relevant departments or construction companies, including ongoing or planned road construction, temporary traffic restrictions, etc. Road construction is one of the potential risk factors for traffic accidents. Including it in the analysis scope can improve prediction accuracy. Through the above detailed steps, multi-source data such as historical traffic data, real-time data, social media data, weather forecast data, and road construction information can be collected to provide a more comprehensive information basis for potential risk analysis of traffic accidents.

[0048] Step 2: Preprocess and extract features of multi-source data;

[0049] In the embodiment of the present invention, the multi-source data is cleaned, outliers are removed, and data conversion is performed to ensure the accuracy and consistency of the data and prepare for subsequent analysis and modeling;

[0050] Exemplarily, first, data cleaning is performed, including removing duplicate data, processing missing values, processing outliers, etc., and then data conversion is performed to convert the date and time into a unified time format to facilitate subsequent time series analysis. At the same time, the data is standardized or normalized. The specific method is as follows: the standardization process is a two-step process. First, the mean is calculated, which can be obtained by summing the eigenvalues ​​and dividing by the number of samples. Finally, the average of the sum of the squares of the differences between each eigenvalue and the mean is calculated, and the square root is taken to obtain the standard deviation. Then, each eigenvalue is standardized by subtracting the mean from the eigenvalue and then dividing by the standard deviation. , get the standardized value, the standardized value has zero mean and unit standard deviation, that is, the mean is 0 and the standard deviation is 1; Normalization processing: calculate the minimum and maximum values ​​of the feature at the same time and normalize the feature values. For each feature value, subtract the minimum value from the feature value, and then divide it by the difference between the maximum and minimum values ​​to get the normalized value. The normalized value ranges from 0 to 1, which represents the ratio of the minimum and maximum values ​​relative to the feature. After the data preprocessing is completed, we need to extract features from historical traffic data and real-time data. These features will be used to build a prediction model to identify potential risk areas for traffic accidents;

[0051] Exemplarily, different attributes of the data are extracted as features. When the location is used as a feature, the location is represented by geocoding or latitude and longitude and converted into numerical features. The time of the traffic accident is used as a feature, and time information such as hours, dates, and days of the week can be extracted and converted into numerical features. Weather information is used as a feature, and weather factors such as rainfall, visibility, and temperature can be included and converted into numerical features. Road types are used as features, such as highways, urban roads, etc., and road types can be encoded as numerical features. Traffic flow and speed are used as features, which can be calculated based on historical traffic data and real-time data. Through data preprocessing and feature extraction, representative and information-rich features can be extracted from the raw data to provide reliable input for subsequent analysis and modeling. These features will be used as input variables of the prediction model to help the system identify potential risk areas for traffic accidents.

[0052] Step 3: Establish prediction model and evaluation;

[0053] In the embodiment of the present invention, machine learning and statistical analysis methods are used to perform model training and optimization on multi-source data;

[0054] Exemplarily, during the training process, historical traffic data is first used as a training set. Each training sample consists of features such as location, time, weather, road type, and traffic flow, and corresponds to a label indicating whether a traffic accident has occurred at the location. The goal is to build a decision tree model through these features, which can accurately predict traffic accidents based on the conditions of the input features. In the process of building the decision tree, the splitting method of each node is determined by the splitting criterion and algorithm. The splitting criterion is used to calculate the information gain or Gini index value of each feature to measure the importance of the feature for classification. According to the result of the criterion calculation, the feature with the maximum information gain or the minimum Gini index is selected as the splitting feature of the current node;

[0055] For example, in the decision tree algorithm, information gain can be used to measure the importance of features for classification. The steps for calculating information gain are as follows: first, the entropy of the data set before division is calculated, and the entropy of the data set is calculated using the proportion of each category. For a binary classification problem, the entropy calculation formula is as follows: Among them, pi represents the probability of category i. At the same time, for each feature, the conditional entropy after division according to the feature is calculated. The conditional entropy represents the entropy of the data set when the feature value is known. The formula for calculating the conditional entropy is as follows: Among them, m is the number of values ​​of feature A, Di represents the sub-dataset when feature A takes value i, and finally the information gain is calculated. The information gain represents the reduction in entropy of the entire system after the feature is divided, that is, the entropy before division minus the conditional entropy. The formula for calculating information gain is as follows: Gain(D, A) = Entropy(D)-H(D\A), select the feature with the largest information gain as the split feature of the current node, and recursively apply the splitting process. We divide the data set into different subsets and continue the splitting operation on each subset until the stopping condition is met. The stopping condition can be that the maximum depth is reached, the number of samples in the node is insufficient, or the splitting criterion no longer improves the prediction ability. In the training process of the decision tree, each leaf node represents a category or decision result. When a new data sample enters the decision tree, it will traverse along the branch of the tree according to the feature value of the sample, and finally reach a leaf node, and make a prediction based on the category or decision result represented by the leaf node;

[0056] Exemplarily, during the model training process, a cross-validation method can be used to make better use of a limited data set. For example, k-fold cross-validation divides the data set into k subsets, selects one of the subsets as the validation set in turn, and the remaining subsets as the training set. Through multiple iterations, the average value of the performance indicators of the model on different validation sets is calculated to obtain a more reliable evaluation result. Once the model has been evaluated and achieves satisfactory performance, it can be applied to real-time data for traffic accident prediction and early warning. Real-time data includes vehicle flow, vehicle speed, weather conditions, etc. By combining with the prediction model, the risk index or probability is calculated instantly, and potential risk areas for traffic accidents are identified. When the risk exceeds the preset threshold, the early warning system will trigger an alarm and remind the driver and traffic management department to take corresponding measures, such as adjusting traffic lights, restricting vehicle traffic or strengthening patrols. Step three involves establishing a prediction model, optimizing model parameters, evaluating model performance, and applying it to real-time data for traffic accident prediction and early warning. This process will help improve traffic safety and provide targeted measures and decision support for drivers and traffic management departments.

[0057] Step 4: Real-time prediction and comparison of traffic data;

[0058] In the embodiment of the present invention, real-time data is input into the established prediction model, and real-time traffic accident prediction is performed, aiming to determine whether there is a potential traffic accident risk area based on the current real-time data, and compare the prediction results with the actual data to verify the accuracy and prediction ability of the model;

[0059] Exemplarily, real-time data is obtained, including vehicle flow, vehicle speed, weather conditions, etc., and then the real-time data is input into the prediction model for prediction. The prediction model uses the model established in step three and uses the real-time data as input features for prediction. The model will calculate the potential traffic accident risk index or probability based on the characteristic value of the real-time data. These indexes or probabilities reflect the potential degree of traffic accidents at different locations at the current moment. The prediction results are compared with the actual data to determine whether there are potential traffic accident risk areas. An early warning threshold is set. When the prediction result exceeds the threshold, it means that there is a high risk of traffic accidents at the location. At this time, the early warning system will trigger an alarm and convey the early warning information to the relevant drivers and traffic management departments through different communication channels, reminding them to take corresponding safety measures and traffic control measures;

[0060] For example, during the comparison process, part of the real-time data can be used as a validation set through the cross-validation method to verify and evaluate the prediction results. By comparing with the actual traffic accident data, the accuracy and predictive ability of the model can be further verified. If the prediction results of the model are highly consistent with the actual data, it means that the model has a high prediction accuracy and can provide reliable traffic accident warning services.

[0061] Step 5: The system issues an early warning based on the prediction results.

[0062] In the embodiment of the present invention, according to the prediction result of step 4, when a location with a risk of traffic accidents is found, a warning message is issued in time to remind the driver and the traffic management department to take corresponding measures. The warning message can be conveyed to relevant personnel in various forms, such as traffic navigation systems, mobile phone applications, electronic billboards, etc.;

[0063] For example, once the prediction model determines that there is a high risk of traffic accidents at a certain location, the system will trigger the early warning mechanism, and the early warning information will be generated and customized according to factors such as location, time, and risk level. For example, through the traffic navigation system, drivers can receive sound prompts or text warnings during navigation, informing them that they are about to enter an area with a risk of traffic accidents, and they are advised to drive carefully or choose an alternative route. Mobile applications can send early warning information to drivers via push notifications so that they can understand and adjust their driving plans in a timely manner. Electronic signs can display relevant warning information on the road to remind drivers to pay attention to safety and abide by traffic rules. Through traffic navigation systems, mobile applications, electronic signs, etc., early warning information is promptly conveyed to drivers and traffic management departments to remind them to pay attention to traffic accident risk areas and take corresponding safety measures.

[0064] Embodiment 2: Embodiment 2 of the present invention provides an abnormal analysis system based on big data. Figure 2 A schematic diagram of the module composition of the big data-based anomaly analysis system provided in the second embodiment of the present invention is shown in FIG. Figure 2 As shown, the system includes:

[0065] Data collection and preprocessing module, used for multi-source data collection and processing;

[0066] Prediction model building and evaluation module, used to build traffic accident models;

[0067] Real-time prediction and warning module, used to predict traffic accidents and issue warnings based on real-time data;

[0068] In some embodiments of the present invention, the data collection and preprocessing module includes:

[0069] Multi-source data collection module, used to collect historical traffic data, real-time data, social media data, weather forecast data, road construction information and other data sources;

[0070] Data cleaning and outlier processing module, used to clean the collected data, remove duplicate, missing and erroneous data, and ensure the accuracy and consistency of the data;

[0071] Feature extraction and conversion module, used to extract features from historical traffic data and real-time data;

[0072] In some embodiments of the present invention, the prediction model building and evaluation module includes:

[0073] Model selection and training set division module, used to select a suitable prediction model and divide historical traffic data into training set and validation set for model training and evaluation;

[0074] The model training and optimization module is used to use the supervised learning method to input the training set into the prediction model for training and optimization;

[0075] The model evaluation and verification module is used to evaluate the performance of the model using the verification set, calculate the accuracy, recall, precision and other evaluation indicators, and verify the predictive ability and stability of the model;

[0076] In some embodiments of the present invention, the real-time prediction and warning module includes:

[0077] The real-time data input and prediction module is used to input real-time data into the established prediction model, conduct real-time traffic accident prediction, and calculate the potential traffic accident risk index or probability;

[0078] The prediction result comparison and judgment module is used to compare the prediction results with the actual data to determine whether there is a potential traffic accident risk area;

[0079] The warning information issuance and communication module is used to generate warning information based on the prediction results and convey it to drivers and traffic management departments through traffic navigation systems, mobile phone applications, electronic billboards, etc.

[0080] 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.

[0081] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An abnormal analysis method based on big data, characterized by: The method comprises the following steps: Step 1: Ensure the comprehensiveness of data through multi-source data collection; Step 2: Preprocess and extract features of multi-source data; Step 3: Establish prediction model and evaluation; Step 4: Real-time prediction and comparison of traffic data; Step 5: The system issues an early warning based on the prediction results.

2. The method for abnormal analysis based on big data according to claim 1, characterized in that: The steps to ensure the comprehensiveness of data through multi-source data collection include: Collect historical traffic data and real-time data; Synthesize information from multiple data sources.

3. The method for abnormal analysis based on big data according to claim 2, characterized in that: The step of integrating information from multiple data sources includes: Obtain weather forecast data through the meteorological department or meteorological data providers. These data include rainfall, visibility, temperature and weather information related to traffic safety. Ensure that accurate weather forecast data is obtained so that it can be correlated with traffic accident risks for analysis. Finally, through cooperation with relevant departments or construction companies, obtain road construction information, including ongoing or planned road construction and temporary traffic restrictions. Road construction is one of the potential risk factors for traffic accidents. Including it in the analysis scope can improve prediction accuracy. Through the above detailed steps, historical traffic data, real-time data, social media data, weather forecast data and road construction information multi-source data can be collected to provide a more comprehensive information basis for potential risk analysis of traffic accidents.

4. The method for abnormal analysis based on big data according to claim 1, characterized in that: The step of preprocessing and feature extraction of multi-source data includes: Perform data preprocessing to ensure data accuracy and consistency; Extract features from historical traffic data and real-time data; The extracted feature data is converted into numerical values.

5. The method for abnormal analysis based on big data according to claim 4, characterized in that: The step of converting the extracted feature data into numerical values ​​comprises: Different attributes of the data are extracted as features. When the location is used as a feature, the location is represented by geocoding or latitude and longitude, and converted into numerical features. The time of the traffic accident is used as a feature, and the hour, date, and day of the week time information can be extracted and converted into numerical features. Weather information can be used as a feature, including rainfall, visibility, and temperature weather factors, and converted into numerical features. Road types are used as features, such as highways and urban roads, and road types can be encoded as numerical features. Traffic flow and speed can be used as features. They can be calculated based on historical traffic data and real-time data. Through data preprocessing and feature extraction, representative and information-rich features can be extracted from the original data to provide reliable input for subsequent analysis and modeling. These features will be used as input variables of the prediction model to help the system identify potential risk areas for traffic accidents.

6. The method for abnormal analysis based on big data according to claim 1, characterized in that: The steps of establishing a prediction model and evaluating include: Use decision tree algorithm to establish traffic accident prediction model; Use supervised learning methods with historical traffic data as input features; Use evaluation metrics to validate the predictive power of the model.

7. The method for abnormal analysis based on big data according to claim 6, characterized in that: The step of using the decision tree algorithm to establish a traffic accident prediction model includes: During the training process, historical traffic data is first used as the training set. Each training sample consists of location, time, weather, road type, and traffic flow characteristics, and corresponds to a label indicating whether a traffic accident has occurred at the location. The goal is to use these features to build a decision tree model that can accurately predict traffic accidents based on the conditions of the input features. In the process of building the decision tree, the splitting criterion and algorithm are used to determine the splitting method of each node. The splitting criterion is used to calculate the information gain or Gini index value of each feature to measure the importance of the feature for classification. According to the results of the criterion calculation, the feature with the maximum information gain or the minimum Gini index is selected as the splitting feature of the current node.

8. The method for abnormal analysis based on big data according to claim 1, characterized in that: The step of real-time prediction and comparison of traffic data includes: Input real-time data into the prediction model to make real-time traffic accident predictions; Compare the predicted results with the actual data to determine the areas of traffic accident risk.

9. The method for abnormal analysis based on big data according to claim 1, characterized in that: The step of the system issuing an early warning according to the prediction result comprises: Once the prediction model determines that a certain location has a high risk of traffic accidents, the system will trigger the early warning mechanism. The early warning information will be generated and customized according to the location, time, and risk level factors. Through the traffic navigation system, drivers can receive voice prompts or text warnings during the navigation process, informing them that they are about to enter an area with a risk of traffic accidents, and are advised to drive carefully or choose an alternative route. Mobile applications can send early warning information to drivers via push notifications so that they can understand and adjust their driving plans in a timely manner. Electronic billboards can display relevant warning information on the road to remind drivers to pay attention to safety and abide by traffic rules. Through the traffic navigation system, mobile applications, and electronic billboards, early warning information can be promptly conveyed to drivers and traffic management departments to remind them to pay attention to traffic accident risk areas and take corresponding safety measures.

10. The abnormal analysis system based on big data is characterized by: The system comprises: Data collection and preprocessing module, used for multi-source data collection and processing; The prediction model establishment and evaluation module is used to establish traffic accident models; the real-time prediction and warning module is used to predict traffic accidents and issue warnings based on real-time data.

Images