Method and device for evaluating district station building based on multi-dimensional data

Abstract

The invention relates to the technical field of site selection optimization, and discloses a district site establishment evaluation method and device based on multi-dimensional data, and the method comprises the steps: constructing a training sample which comprises a charging station identifier, charging station establishment time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, determining a sample label based on the difference between the charging amounts before and after the building of the charging station in the training sample; based on the training samples, statistical features are constructed, a district station building recommendation model is trained based on the statistical features, and the statistical features comprise the charging amount and the travel times of the district where the charging station is located in the training samples; and generating a station building evaluation result of the station building candidate districts based on the station building candidate district data and the trained district station building recommendation model. According to the method, the working efficiency of site selection is improved, multi-dimensional data such as charging quantity requirements and traffic convenience can be considered, model interpretability can be given, and region screening can be carried out more reasonably and more efficiently.

Description

Technical Field

[0001] The present invention relates to the technical field of site selection optimization, and in particular to a method and device for evaluating site construction in an area based on multi-dimensional data. Background Art

[0002] As the number of new energy vehicles in my country continues to increase, the proportion of new energy vehicles in the total number of vehicles is also increasing. At present, the ratio of public charging piles to new energy vehicles in my country is greater than 7. Although the ratio of public charging piles is generally on a downward trend, there is still a large gap in public charging piles.

[0003] The construction of charging stations often requires a large amount of equipment and financial costs. A wrong site selection decision means that the investment cannot get a reasonable return, which will result in a waste of manpower and material resources. Summary of the invention

[0004] In view of this, the present invention provides a method and device for evaluating charging station construction in an area based on multi-dimensional data, so as to solve the problem of waste of manpower and material resources caused by wrong decision-making in site selection of charging stations.

[0005] In a first aspect, the present invention provides a method for evaluating area station construction based on multi-dimensional data, the method comprising: constructing a training sample, the training sample comprising a charging station identifier, a charging station construction time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on the difference in the charging amount before and after the charging station construction in the training sample; constructing statistical features based on the training sample, and training an area station construction recommendation model based on the statistical features, the statistical features comprising the charging amount and the number of trips in the area where the charging station is located in the training sample; generating an evaluation result of the candidate area station construction based on the candidate area station construction data and the trained area station construction recommendation model.

[0006] In an optional embodiment, constructing statistical features based on the training samples includes at least one of the following: constructing the statistical features based on the number of candidate areas for establishing charging stations in the training samples; constructing the statistical features based on the number of roads in the areas where charging stations are located in the training samples.

[0007] In an optional embodiment, the training of the area site recommendation model based on the statistical features includes: initializing a decision tree; iteratively training the decision tree based on the statistical features until the end conditions of the area site recommendation model are met, thereby obtaining a trained area site recommendation model; wherein, a new decision tree is constructed in each iteration, and a previous decision tree of the new decision tree is corrected based on the new decision tree; during the iterative training process, the loss function is minimized; for the decision tree or the new decision tree, a split point is selected based on the split gain, wherein the split point is determined based on the leaf node with the largest split gain, and based on the split gain, the increase in the purity of the leaf node after the split is evaluated.

[0008] In an optional embodiment, the construction of training samples includes: obtaining charging station information that has been established within a preset time range, the charging station information including the charging station identifier, the charging station location information, and the charging station establishment time; determining the H3 index corresponding to the location information; if the difference is greater than or equal to a preset threshold, the sample label is a positive sample; if the difference is less than the preset threshold, the sample label is a negative sample; the training sample constructed includes the H3 index, the charging station establishment time and the sample label.

[0009] In an optional implementation, the generating of the site construction evaluation result of the candidate site construction area based on the candidate site construction area data and the trained site construction recommendation model for the area includes: inputting the candidate site construction area data into the trained site construction recommendation model for the area, generating a recommendation index corresponding to the candidate site construction area, if the recommendation index of the target candidate site construction area is greater than a preset recommendation threshold, recommending to build a charging station in the target candidate site construction area; wherein the candidate site construction area data includes the location information of the candidate site construction area and the estimated site construction time.

[0010] In an optional implementation, the method further includes: determining a characteristic value of the area site establishment recommendation model based on a mean value of the statistical feature; and calculating the splitting gain based on the characteristic value.

[0011] In a second aspect, the present invention provides a device for evaluating area station construction based on multi-dimensional data, the device comprising: a sample construction module, used to construct a training sample, the training sample comprising a charging station identifier, a charging station construction time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on the difference in the charging amount before and after the charging station construction in the training sample; a model construction module, used to construct statistical features based on the training sample, and train an area station construction recommendation model based on the statistical features, the statistical features comprising the charging amount and the number of trips in the area where the charging station is located in the training sample; an evaluation module, used to generate an evaluation result of the candidate area station construction based on the candidate area station construction data and the trained area station construction recommendation model.

[0012] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method for area site evaluation based on multi-dimensional data according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0013] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for evaluating area site establishment based on multi-dimensional data according to the first aspect or any corresponding embodiment thereof.

[0014] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions for causing a computer to execute the method for evaluating area site establishment based on multi-dimensional data according to the first aspect or any corresponding embodiment thereof.

[0015] Based on the historical charging station construction data, a district station construction recommendation model is constructed. The district station construction recommendation model can predict the possible effects of establishing a charging station based on the statistical characteristics of the candidate districts for station construction, and provide decision-making support for the construction of charging stations. Compared with the method of using manual experience to screen the station construction area, the efficiency of site selection is improved, and multi-dimensional data such as charging demand and traffic convenience can be taken into account. The model can be explained, and regional screening can be more reasonable and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A schematic diagram showing a flow chart of a method for evaluating site construction in a region based on multi-dimensional data according to an embodiment of the present invention;

[0018] Figure 2 A schematic diagram showing the structure of a device for evaluating site construction in a region based on multi-dimensional data according to an embodiment of the present invention is shown;

[0019] Figure 3 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0021] There are two methods for deciding the location of charging stations in related technologies:

[0022] The first is that operators go to the intended locations for offline inspections to gain a preliminary understanding of the data and distribution of new energy vehicles in the area. However, this is only a short-term perceptual understanding and cannot provide a complete inspection from the dimensions of daytime, nighttime, seasons, holidays, and surrounding environment. Decisions rely on subjective judgment.

[0023] Second, some operators use site selection products developed by map vendors or some software. These site selection products are generally developed for traditional chain formats and have the functions of surrounding commercial competition, road traffic flow and population portrait. Charging station site selection is similar to normal chain format site selection. The site selection platform in related technologies can solve problems such as site, economic development and population consumption capacity in the site selection process, but it cannot directly reflect the charging demand of new energy vehicles.

[0024] According to an embodiment of the present invention, a method embodiment of area site evaluation based on multi-dimensional data is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] In this embodiment, a method for evaluating site construction in a region based on multi-dimensional data is provided, which can be used in mobile terminals, such as mobile phones, tablet computers, desktop computers, or servers, etc. Figure 1A schematic diagram of a process flow of a method for evaluating a site construction area based on multi-dimensional data according to an embodiment of the present invention is shown. Figure 1 As shown, the process includes the following steps:

[0026] Step S101, constructing a training sample, the training sample includes a charging station identifier, a charging station establishment time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on the difference in charging amount before and after the charging station is established in the training sample.

[0027] In this step, a unique identifier can be assigned to each charging station so that different charging stations can be accurately identified and distinguished in subsequent data processing and model training. The establishment time of each charging station is recorded. In this way, the difference in charging capacity before and after the charging station is established can be calculated later to determine the sample label. The establishment time should be accurate to a sufficient granularity to accurately reflect the change in charging capacity. The establishment time can be accurate to the month.

[0028] The sample label is determined based on the difference in charging volume before and after the charging station is built. Specifically, the difference between the charging volume within a certain period of time (such as three months) after the charging station is built and the charging volume in the same period of time before the construction can be calculated. This difference reflects the effect of the charging station on the improvement of charging volume, so it can be used as a sample label to train the model.

[0029] Step S102, constructing statistical features based on the training samples, and training an area station recommendation model based on the statistical features, wherein the statistical features include the charging amount and the number of trips in the area where the charging station in the training samples is located.

[0030] In this step, the charging stations are divided into different areas according to geographical location, administrative divisions or other relevant standards. Each area contains a certain number of charging stations. For each area, the total charging amount of all charging piles in the area in the training sample within a certain period of time is counted. This feature can reflect the charging demand level and the use of electric vehicles in the area. Similarly, for each area, the total number of trips of electric vehicles in the area in the training sample within a certain period of time is counted. This feature can reflect the traffic convenience, electric vehicle activity and mobility of the area.

[0031] Recommendation models play a vital role in today's data analysis and machine learning fields. They are mainly used to predict users' interest in specific items, thereby optimizing the performance of recommendation systems. The core goal of recommendation models is to predict users' interest in items they have not been exposed to by analyzing their historical behavior data, and then provide users with personalized recommendations.

[0032] In this step, the area site recommendation model is used to evaluate and screen candidate areas for site construction, and select those areas with probabilities greater than the preset recommendation threshold, replacing the traditional manual experience screening method. At the same time, it gives the feature importance of factors such as charging demand and traffic convenience, providing customers with reasonable explainability. You can choose the Light Gradient Boosting Machine (LightGBM) as the area site recommendation model.

[0033] Step S103: generating a site construction evaluation result of the candidate site construction area based on the site construction candidate area data and the trained area site construction recommendation model.

[0034] In this step, the candidate area data includes name, type, longitude, latitude, province, city and expected month of commissioning. According to the statistical features used in training the area recommendation model, the corresponding features are extracted from the candidate area data. The candidate area data can be aggregated to obtain input features. In geographic information systems, a POI can be a house, a shop, a mailbox or a bus stop.

[0035] Based on the input features and the trained area site recommendation model, the site evaluation results of the candidate site areas are obtained. The evaluation results are used to characterize the site recommendation index of the candidate site areas.

[0036] The method for evaluating district station construction based on multi-dimensional data provided in this embodiment constructs a district station construction recommendation model based on historical charging station construction data. The district station construction recommendation model can predict the possible effects of establishing a charging station based on the statistical characteristics of candidate districts for station construction, and provide decision-making support for charging station construction. Compared with the method of using manual experience to screen station construction areas, the efficiency of site selection is improved, and multi-dimensional data such as charging demand and traffic convenience can be taken into account, the model can be explained, and regional screening can be more reasonable and efficient.

[0037] In some optional embodiments, statistical features are constructed based on the training samples, including at least one of the following: constructing statistical features based on the number of candidate areas for station construction in the area where the charging stations are located in the training samples; constructing statistical features based on the number of roads in the area where the charging stations are located in the training samples.

[0038] In this embodiment, the number of candidate areas for station construction reflects the number of potential locations for charging station construction in the area, which can be used to evaluate the competition situation of operators in the area. The number of roads can be obtained based on the aggregation of road network data. The road network data may include the road name, type, starting latitude and longitude, province, and month of commissioning.

[0039] The number of roads reflects the traffic density and accessibility of the area, which is closely related to the driving and charging needs of electric vehicles. The more roads there are, the greater the traffic volume in the area, and the charging demand of electric vehicles may also increase accordingly.

[0040] Statistical features can also be constructed based on the traffic flow data of the area where the charging station is located in the training sample. Traffic durian data includes H3 index, province, city, number of trips and month.

[0041] In this way, during the training process of the area station recommendation model, multi-dimensional factors are taken into account, such as the demand for charging capacity, the competition situation of surrounding operators, traffic convenience, business environment, etc., to improve the model effect and provide better explainability. At the same time, vehicle usage behavior reflects the user's usage preferences. User characteristics such as vehicle parking and usage frequency determine the demand for power, and this part of data is precisely what map service providers do not have.

[0042] In some optional embodiments, the area site recommendation model is trained based on statistical features, including: initializing a decision tree; iteratively training the decision tree based on statistical features until the end conditions of the area site recommendation model are met, thereby obtaining a trained area site recommendation model; wherein, each iteration constructs a new decision tree, and corrects the previous decision tree based on the new decision tree; during the iterative training process, the loss function is minimized; for the decision tree or the new decision tree, a split point is selected based on the split gain, wherein the split point is determined based on the leaf node with the largest split gain, and based on the split gain, the increase in the purity of the leaf node after the split is evaluated.

[0043] In this embodiment, the initialized decision tree as a basic model can be relatively simple, for example, it can be a constant model. More decision trees are constructed in turn in an iterative manner. Each newly constructed decision tree attempts to correct the error of the previous decision tree, that is, to reduce the prediction error. In each iteration, the area site recommendation model is optimized according to the gradient information. The gradient represents the derivative of the prediction error of the area site recommendation model at the current data point, which indicates how to adjust the model parameters to minimize the loss function.

[0044] For each decision tree, the optimal features and split points are selected based on the split gain. The split gain measures the improvement in the purity of the child nodes after the split. LightGBM uses a histogram algorithm to speed up this process, discretizing continuous feature values ​​into finite histogram buckets, thereby reducing the amount of calculation. LightGBM adopts a leaf node growth strategy (Leaf-wiseGrowth), that is, in each iteration, the leaf node with the largest gain is selected for splitting. This strategy enables the model to fit the training data more quickly.

[0045] The key parameters params in the recommended model for regional site building are as follows:

[0046]

[0047] Among them, n_estimators is the number of decision trees to be built, that is, the number of iterations or the number of boosting rounds; max_depth is the maximum depth of the tree; num_leaves is the number of leaf nodes in a tree; min_child_samples is the minimum number of data on a leaf node; colsample_bytree is the proportion of random sampling of features when building a decision tree; learning_rate is the contribution of each decision tree to the final prediction; subsample is the proportion of random sampling of samples when training each tree.

[0048] After training is completed, the new data is predicted using the region site recommendation model. LightGBM outputs the site recommendation index.

[0049] During the training of the recommended model for regional site building, all positive and negative samples are retained without sampling. All samples are randomly divided into training sets and test sets according to a preset ratio, for example, a ratio of 7:3 can be set for model training and testing.

[0050] The decision tree model has excellent interpretability. Each node represents a feature, each edge represents a decision rule, and the final leaf node is the prediction result. This intuitive tree structure makes it easy for non-professionals to understand how the model works and the logic behind the prediction results. In addition, the decision tree model can easily handle multiple types of data, including numerical data and categorical data, and can be directly used for model training without complex preprocessing or data conversion, which is more flexible.

[0051] In some alternative embodiments, constructing training samples includes: obtaining charging station information established within a preset time range, where the charging station information includes a charging station identifier, charging station location information, and charging station construction time; determining the H3 index corresponding to the location information; if the difference is greater than or equal to a preset threshold, the sample label is a positive sample; if the difference is less than the preset threshold, the sample label is a negative sample; the constructed training samples include the H3 index, charging station construction time, and sample label.

[0052] H3 is a geospatial indexing system that divides the Earth's surface into a series of hexagonal grids and also supports some pentagonal grids. The H3 indexing system numbers and manages these grids in a hierarchical structure, and each grid has a unique index code. The H3 indexing system balances the uniformity of area division and computational efficiency by using the geometric shape of hexagons. The H3 indexing system has 16 levels, from 0 to 15, and the higher the level, the smaller the hexagonal grid. The 8th-level index of H3 is a medium-grained grid, covering an area of about 5 square kilometers, which is suitable for geospatial analysis and data processing at the city level.

[0053] In this embodiment, the station information table includes a charging station identifier, charging station name, charging station construction time, number of AC chargers, number of DC chargers, longitude, latitude, province, and city. The stations in the station information table whose construction time belongs to a certain period are used as sampling targets. The longitude and latitude are uniformly converted into WGS-84 coordinates. WGS-84 is an internationally widely used geographic coordinate system that is defined based on the shape and size of the Earth and provides a unified geographic reference framework for the world. Functions in the H3 library (e.g., Uber's H3 Python library) can be used to convert WGS-84 coordinates into H3 indexes. Based on the above data, find all sets of <H3 index, construction month> as preliminary training samples.

[0054] Based on the preliminary training samples constructed in the previous step, for each sample, by associating the charging information data table with the H3 index, the average charging volume A in the 3 months after the construction month and the charging volume B in the 1 month before the construction month can be obtained. If A - B >= a certain threshold, it is marked as 1 and used as a positive sample; otherwise, it is marked as 0 and used as a negative sample. The obtained training samples include the H3 index, charging station construction time, and sample label. Among them, the charging information data table includes: H3 index, charging volume in the area where the charging station is located, identity, and month.

[0055] Based on the H3 area where charging stations already exist as training samples, the difference in charging meter display before and after the construction of the H3 area is used as training labels, so that the model can correctly learn the recommendation index of the H3 area. Machine learning modeling is carried out based on multi-dimensional data, and the model characteristics of the H3 area are obtained based on factors such as charging demand, operator competition, traffic convenience, and business environment. LightGBM is used to build a recommendation model, which can achieve good differentiation effect.

[0056] Obtaining the information of charging stations that have been established within the preset time range ensures the timeliness and relevance of the training samples; the charging station information includes the charging station identifier, charging station location information, and charging station construction time, which provides rich features for subsequent sample construction. Determining the H3 index corresponding to the location information can convert the geographic location information of different data sources into a numerical form that the model can process, which is convenient for subsequent feature processing and model training. Determining the sample label based on the comparison result between the difference and the preset threshold can divide the charging station information into positive samples and negative samples, providing a clear supervision signal for the model, which helps the model learn how to distinguish different situations. Since the training samples are generated based on actual data, they have high authenticity and reliability, which helps to improve the generalization ability of the model.

[0057] In some optional embodiments, based on the candidate area data for station construction and a trained area site construction recommendation model, a site construction evaluation result for the candidate area for station construction is generated, including: inputting the candidate area data for station construction into the trained area site construction recommendation model to generate a recommendation index corresponding to the candidate area for station construction; if the recommendation index of the target candidate area for station construction is greater than a preset recommendation threshold, it is recommended to establish a charging station in the target candidate area for station construction; wherein the candidate area data for station construction includes location information of the candidate area for station construction and an estimated time for station construction.

[0058] In this embodiment, the preset recommendation threshold can be determined according to the actual situation and business needs, and is used to determine whether the recommendation index reaches the standard for recommending site establishment. Based on the trained area site establishment recommendation model, the plots with recommendation scores greater than the preset recommendation threshold are screened out. This method replaces the traditional manual experience screening method, takes into account multi-dimensional factors, and provides model interpretability, which can perform regional screening more reasonably and efficiently.

[0059] In some optional implementations, the aforementioned method for evaluating area site building based on multi-dimensional data further includes: determining a characteristic value of an area site building recommendation model based on a mean value of statistical characteristics; and calculating a split gain based on the characteristic value.

[0060] In this embodiment, for each training sample, the feature mean of the H3 index corresponding to the training sample in several months before the station construction month is counted, and the feature mean includes the charging amount of the area where the charging station is located, the total number of candidate areas for station construction, the number of area trips, and the number of roads in the area. The feature mean is used as the feature value of the area station construction recommendation model.

[0061] When building a decision tree, the sharding recommendation model will traverse all possible splitting points for each feature and calculate the splitting gain to determine the best splitting method. The splitting gain is usually calculated by comparing a certain indicator before and after the split, such as information gain. For example, when calculating information gain, the sharding recommendation model considers the change in information entropy after each feature value divides the data set into different subsets. Different combinations and distributions of feature values ​​will lead to different changes in information entropy, which in turn affects the size of the splitting gain. At each node of the decision tree, the sharding recommendation model will select a feature and its corresponding splitting point for splitting. This selection process is usually based on the size of the splitting gain. The feature value indirectly affects the selection of the feature by affecting the calculation result of the splitting gain. Features with larger splitting gains are more likely to be selected for splitting nodes.

[0062] By determining the feature value by taking the mean of the statistical features, we can ensure that the area site recommendation model is based on a large amount of objective data to make decisions, rather than relying on subjective judgment or experience. This can more accurately capture the key factors of area site construction, thereby generating more reliable recommendation results.

[0063] In some optional implementations, the area under the ROC curve (Area Under Curve, referred to as AUC) can be used to evaluate the ability of the regional site recommendation model to evaluate the probability of site establishment. The ROC curve is generally above the straight line y=x, so the value range of AUC is between 0.5 and 1. The closer the AUC is to 1.0, the higher the authenticity of the detection method; when it is equal to 0.5, the authenticity is the lowest and there is no application value.

[0064] In this way, we can not only quantify the performance of the regional site recommendation model, but also adapt to the situation of sample imbalance and do not rely on specific classification thresholds. After testing and verification, LightGBM can obtain a good AUC.

[0065] In some optional implementations, the evaluation results of the candidate areas for site building can be updated on a monthly basis. Based on the multi-dimensional basic data of the H3 area generated monthly, feature calculation is performed, and then the area site building recommendation model score prediction is performed based on the features, and the recommendation scores of all H3 areas in the site building evaluation results are updated. According to customer needs, a suitable preset recommendation threshold is selected, and H3 areas greater than the preset recommendation threshold are screened and recorded.

[0066] Traditional site selection refers to data such as POI, AOI, population, and traffic volume, which reflect the environmental information around the point. Its update cycle is recorded in years. The method of regional site evaluation based on multi-dimensional data in the present invention can accurately grasp the rhythm of the market and discover market opportunities in a timely manner through higher-frequency data updates.

[0067] It is possible to conduct website building evaluation for all H3 areas across the country, perform distributed computing of multi-dimensional features of all H3 areas, and conduct comprehensive prediction and scoring through the regional website building recommendation model.

[0068] The method for evaluating area site construction based on multi-dimensional data provided by the present invention provides an innovative solution for the scientific planning and efficient construction of new energy vehicle charging infrastructure by unifying data, taking into account multi-dimensional recommendation factors, and screening candidate plots.

[0069] In this embodiment, a device for evaluating the site construction in a region based on multi-dimensional data is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0070] This embodiment provides a device for evaluating site construction in a region based on multi-dimensional data. Figure 2 The schematic diagram of the structure of the device for evaluating the site construction in a region based on multi-dimensional data according to an embodiment of the present invention is shown as follows: Figure 2 As shown, including:

[0071] The sample construction module 501 is used to construct a training sample, which includes a charging station identifier, a charging station establishment time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on the difference in charging amount before and after the charging station is established in the training sample.

[0072] The model building module 502 is used to build statistical features based on the training samples, and train the area station building recommendation model based on the statistical features. The statistical features include the charging amount and the number of trips in the area where the charging station in the training samples is located.

[0073] The evaluation module 503 is used to generate a site construction evaluation result of the candidate site construction area based on the site construction candidate area data and the trained area site construction recommendation model.

[0074] In some optional implementations, the model building module 502 includes:

[0075] The first unit of the model construction is used to construct statistical features based on the number of candidate areas for station construction in the areas where the charging stations are located in the training samples; and to construct statistical features based on the number of roads in the areas where the charging stations are located in the training samples.

[0076] In some optional implementations, the model building module 502 further includes:

[0077] The second unit of model construction is used to initialize the decision tree; iteratively train the decision tree based on statistical characteristics until the end conditions of the area site recommendation model are met, and a trained area site recommendation model is obtained; wherein, a new decision tree is constructed in each iteration, and the previous decision tree of the new decision tree is corrected based on the new decision tree; during the iterative training process, the loss function is minimized; for the decision tree or the new decision tree, the split point is selected based on the split gain, wherein the split point is determined based on the leaf node with the largest split gain, and based on the split gain, the increase in the purity of the leaf node after the split is evaluated.

[0078] In some optional implementations, the sample construction module 501 includes:

[0079] The sample constructs the first unit, which is used to obtain the information of charging stations that have been established within a preset time range. The charging station information includes the charging station identifier, the charging station location information, and the charging station establishment time; determine the H3 index corresponding to the location information; if the difference is greater than or equal to the preset threshold, the sample label is a positive sample; if the difference is less than the preset threshold, the sample label is a negative sample; the constructed training sample includes the H3 index, the charging station establishment time and the sample label.

[0080] In some optional implementations, the evaluation module 503 includes:

[0081] The evaluation unit is used to input the candidate area data of station construction into the trained area station construction recommendation model, generate a recommendation index corresponding to the candidate area of ​​station construction, and if the recommendation index of the target candidate area of ​​station construction is greater than a preset recommendation threshold, recommend to build a charging station in the target candidate area of ​​station construction; wherein the candidate area data of station construction includes the location information of the candidate area of ​​station construction and the estimated time of station construction.

[0082] In some optional embodiments, the aforementioned device further includes:

[0083] The feature construction module is used to determine the feature value of the area site establishment recommendation model based on the mean value of the statistical feature; and calculate the split gain based on the feature value.

[0084] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0085] The device for evaluating area site construction based on multi-dimensional data in this embodiment is presented in the form of a functional unit, where the unit refers to an application specific integrated circuit (ASIC) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0086] The embodiment of the present invention also provides a computer device having the above Figure 2 The device shown is for evaluating area site establishment based on multi-dimensional data.

[0087] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 3 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display graphical information of a graphical user interface on an external input / output device (such as a display device coupled to an interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 3 A processor 10 is taken as an example.

[0088] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0089] The aforementioned memory 20 stores instructions executable by at least one processor 10, so that the aforementioned at least one processor 10 executes the method shown in the above embodiment.

[0090] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0091] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0092] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0093] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (such as a light emitting diode) and a tactile feedback device (such as a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0094] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0095] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0096] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for evaluating regional site construction based on multi-dimensional data, characterized in that: The method comprises: Constructing a training sample, the training sample comprising a charging station identifier, a charging station establishment time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on a difference in charging amount before and after the charging station is established in the training sample; Based on the training samples, construct statistical features, and train a recommendation model for building a charging station in an area based on the statistical features, wherein the statistical features include the charging amount and the number of trips in the area where the charging station in the training samples is located; Based on the candidate site construction area data and the trained area site construction recommendation model, a site construction evaluation result of the candidate site construction area is generated.

2. The method according to claim 1, characterized in that The constructing of statistical features based on the training samples includes at least one of the following: Constructing the statistical features based on the number of candidate areas for station construction in the areas where charging stations are located in the training samples; The statistical feature is constructed based on the number of roads in the area where the charging stations are located in the training sample.

3. The method according to claim 2, characterized in that The training of the area site building recommendation model based on the statistical features includes: Initialize the decision tree; Iteratively training the decision tree based on the statistical features until an end condition of the area site building recommendation model is met, thereby obtaining a trained area site building recommendation model; Wherein, a new decision tree is constructed in each iteration, and a decision tree before the new decision tree is corrected based on the new decision tree; During the iterative training process, minimizing the loss function; For the decision tree or the new decision tree, a split point is selected based on a split gain, wherein the split point is determined based on a leaf node with the largest split gain, and based on the split gain, an increase in purity of the leaf node after the split is evaluated.

4. The method according to claim 1, characterized in that The constructing of training samples comprises: Acquire the charging station information that has been established within a preset time range, the charging station information including the charging station identifier, the charging station location information, and the charging station establishment time; Determine an H3 index corresponding to the location information; If the difference is greater than or equal to a preset threshold, the sample label is a positive sample; If the difference is less than the preset threshold, the sample label is a negative sample; The training samples constructed include the H3 index, the charging station establishment time and the sample label.

5. The method according to claim 1, characterized in that The generating of the site building evaluation result of the candidate site building area based on the site building candidate area data and the trained area site building recommendation model includes: Input the candidate area data into the trained area recommendation model to generate a recommendation index corresponding to the candidate area. If the recommendation index of the target candidate area is greater than a preset recommendation threshold, it is recommended to build a charging station in the target candidate area. The candidate site building area data includes the candidate site building area location information and the estimated site building time.

6. The method according to claim 3, characterized in that The method further comprises: Determining a characteristic value of the area site establishment recommendation model based on the mean value of the statistical feature; Based on the eigenvalues, the splitting gains are calculated.

7. A device for evaluating site construction in a region based on multi-dimensional data, characterized in that: The device comprises: A sample construction module, used to construct a training sample, wherein the training sample includes a charging station identifier, a charging station establishment time corresponding to the charging station identifier, and a sample label corresponding to the charging station identifier, wherein the sample label is determined based on a difference in charging amount before and after the charging station is established in the training sample; A model building module, used to build statistical features based on the training samples, and train a district station building recommendation model based on the statistical features, wherein the statistical features include the charging amount and the number of trips in the district where the charging station in the training samples is located; The evaluation module is used to generate a site construction evaluation result of the candidate site construction area based on the candidate site construction area data and the trained area site construction recommendation model.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for evaluating area site establishment based on multi-dimensional data according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for evaluating site construction in an area based on multi-dimensional data according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the method for evaluating site construction in an area based on multi-dimensional data according to any one of claims 1 to 6.

Images