Mineral prediction and exploration risk evaluation method based on Bayesian deep learning

Abstract

The invention provides a mineral product prediction and exploration risk evaluation method based on Bayesian deep learning, and relates to the technical field of mineral product exploration, and the method comprises the steps: constructing a multi-source geological data set, and carrying out the preprocessing of the multi-source geological data set, and obtaining an ore deposit and ore control elements; the method comprises the following steps: constructing a training data set through data preprocessing, and further constructing a Bayesian neural network training model based on Metropolis-Hastings sampling; and based on the training model, inputting all ore control element data sets, calculating a probability mean value, accidental uncertainty estimation and cognitive uncertainty estimation, drawing a prediction probability graph, an accidental uncertainty graph and a cognitive uncertainty graph, and performing ore deposit exploration risk and prospecting potential evaluation. According to the method, a sampling method based on Monte Carlo Metropolis-Hastings is adopted, more reliable metallogenic probability mapping and uncertainty estimation are provided, and cognitive uncertainty and accidental uncertainty can be effectively distinguished.

Description

Technical Field

[0001] The present application relates to the field of mineral exploration technology, and in particular to a Bayesian deep learning method for mineral prediction and exploration risk assessment. Background Art

[0002] With the decrease in the number of outcrop deposits and easily discovered deposits, there is an urgent need for advanced, reliable and easy-to-interpret methods to extract and integrate deep-level mineralization information characteristics and find potential prospecting targets. However, the mineralization process is a complex physical and chemical process. Due to the variability of geological phenomena, the complexity of ore deposit genesis, the hidden nature of ore-controlling factors, the multi-solution nature of prospecting information, the uneven distribution of exploration data, the incompleteness of data collection, and the cognitive limitations of describing geological body properties and geological processes, the prediction results of ore deposits are uncertain.

[0003] Although deep learning technology has very superior performance in the intelligent prediction and evaluation of mineral resources, it also has serious defects. For example, machine learning itself is a deterministic model that cannot estimate the uncertainty of the prediction results, which loses its guiding significance for mineral exploration risk assessment. In addition, the machine learning model is not robust enough, and even a slight change in weights may affect the prediction results. Mineralization prediction is a task of making optimal decisions under a variety of uncertain conditions. Scientifically understanding these uncertainties is of great significance for reducing the uncertainty of mineral prediction and the risk of mineral deposit exploration. Summary of the invention

[0004] The purpose of the present invention is to provide a Bayesian deep learning method for mineral prediction and exploration risk assessment in order to solve the problem of low prediction accuracy in existing prospecting target area prediction methods.

[0005] The above-mentioned purpose of the present application is achieved through the following technical solutions: S1: Construct a multi-source geological data set and preprocess it to obtain ore deposits and ore-controlling elements, and construct an ore-controlling element data set; S2: Construct a Bayesian neural network training model based on Metropolis-Hastings sampling; S3: Input the mining control factor data set into the Bayesian neural network training model to obtain the probability mean, aleatory uncertainty estimation, and epistemic uncertainty estimation; S4: Through the probability mean, aleatory uncertainty estimation and epistemic uncertainty estimation, the prediction probability map, aleatory uncertainty map and epistemic uncertainty map are drawn to evaluate the mineral exploration risk and prospecting potential.

[0006] Optionally, step S1 includes: Multi-source geological data sets include: geological data, geochemical data, geographic data, and remote sensing data; Preprocessing includes: missing value processing, outlier processing and noise processing; The preprocessed multi-source address data set is screened to obtain mineral deposits and mineral-controlling elements.

[0007] Optionally, step S2 includes: Set the number of network layers, number of neurons in each layer, and activation function of the Bayesian neural network training model ; Set the prior distribution of weights and biases of the Bayesian neural network training model as follows: Taking the ore deposit and ore-controlling factors as input data, the input data of the Bayesian neural network training model is defined as , the output label is , the training data is ; The neural network parameters are The probability distribution of the predicted output of the Bayesian neural network training model is expressed as follows:

[0008]

[0009]

[0010] In the formula, Output predictions for the Bayesian neural network training model, is the activation function, They are the first The weight matrix and bias vector of the layer.

[0011] Optionally, step S2 further includes: Set the likelihood function of the Bayesian neural network training model. The specific steps are as follows: The presence of mineral deposits is marked as 1, and the absence of mineral deposits is marked as 0. Each sample belongs to one of the two categories; Assume that each sample Tags The value is 0 or 1, and the predicted probability output by the Bayesian neural network training model Representation sample The probability of belonging to category 1; The likelihood function for the binomial distribution is defined as:

[0012]

[0013] in, For sample The true label of express The probability of taking 0 is 0, The probability of taking 1 is ; The probability of taking 0 is , The probability of taking 1 is 0; Convert the likelihood function of the binomial distribution to the log-likelihood function:

[0014] Taking the negative log likelihood, we get the negative log likelihood, which is the loss function ,as follows: .

[0015] Optionally, step S3 includes: S31: Setting the initial neural network parameters for the Bayesian neural network training model ; Select Proposal Distribution , specifically: using Gaussian distribution as a prior distribution;

[0016] in, is the covariance matrix, is a diagonal matrix; represents the Gaussian distribution function; represents the generated parameter samples; Indicates The parameters of the iterations; S32: Perform Metropolis-Hastings sampling to generate parameter samples; For each iterative sampling process, the generated parameter samples are as follows:

[0017]

[0018] in is the total number of iterative sampling processes; Calculating the acceptance rate ,as follows:

[0019] in, Indicated in the parameter Next, label data In the conditions Likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Indicates that the sampling parameters were obtained in the previous iteration Based on the label data In the conditions Likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Determine whether to accept the parameter sample. The specific steps are as follows: Generate a uniformly distributed random number ,like , then accept ,Right now:

[0020] like , then reject ,Right now:

[0021] S33: Approximate the posterior probability distribution by generating parameter samples :

[0022] in, represents training data; Indicates the given parameter Under these conditions, the observed data The likelihood function of ; Representation parameters The prior probability of S34: Save the parameter sample set obtained by sampling ; S35: Perform model prediction using a set of parameter samples, and calculate the probability mean of the prediction probability and the estimate of the aleatory uncertainty; S36: Calculate the variance of the predicted probability under different parameter samples to obtain the epistemic uncertainty estimate.

[0023] Optionally, step S35 includes: The steps to calculate the probability mean are as follows: For new input data , using all sampled parameter samples Perform forward propagation to obtain a series of predicted probabilities , the probability mean after M samplings for:

[0024] The calculation steps of the aleatory uncertainty estimate are as follows: By predicting the variance at the output layer of the Bayesian neural network training model , and obtain the aleatory uncertainty estimate ,as follows:

[0025] Among them, the variance Reflect input data The noise itself; Represents the input data The probability of getting Represents the input data The average probability obtained.

[0026] Optionally, step S36 includes: Calculate the variance of the predicted probability under different parameter samples:

[0027] in, represents the parameter sample of the tth sampling; Representing probability Average.

[0028] Optionally, step S4 includes: Draw a predicted probability map based on the probability mean to form a mineral potential distribution map; Create aleatoric uncertainty maps based on aleatoric uncertainty estimates to identify areas with high data noise; Draw epistemic uncertainty maps based on epistemic uncertainty estimates to identify areas where knowledge is insufficient or data is scarce; Combined with mineral potential distribution maps, areas with high data noise and areas with insufficient knowledge or scarce data, it assists in mineral exploration decision-making, determines priority exploration areas and areas that require further data collection, and completes mineral exploration risk and prospecting potential evaluation.

[0029] An electronic device includes a processor, a memory, a user interface and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes a Bayesian deep learning method for mineral prediction and exploration risk assessment.

[0030] A computer-readable storage medium stores instructions, which, when executed, execute a Bayesian deep learning method for mineral prediction and exploration risk assessment.

[0031] The beneficial effects of the technical solution provided by this application are: 1. Adopt the Monte Carlo-based Metropolis-Hastings method, approximate the true posterior probability distribution through multiple sampling, and then estimate the model parameters and related uncertainties, so as to provide a robust uncertainty estimate for mineral prediction. Specifically, it includes: under the Bayesian framework, the neural network structure and prior distribution are constructed based on the knowledge of domain experts and the characteristics of the input data; based on Bayesian reasoning, the prior probability is combined with the likelihood probability to construct the posterior distribution function; based on the Metropolis-Hastings sampling results, the probability mean, accidental uncertainty and cognitive uncertainty are calculated; finally, the probability mean map, accidental uncertainty map and cognitive uncertainty map are drawn and interpreted. The technical solution of this application integrates deep learning technology and Bayesian reasoning ideas to achieve the measurement and simulation of the uncertainty of the prediction results, and provide support for the risk assessment of mineral exploration.

[0032] 2. The hidden layer in the Bayesian neural network training model structure adopts three fully connected layers. The prior distribution is defined as a normal distribution to simplify the calculation complexity of the model. The likelihood function is defined as a binomial distribution to maintain consistency with the prediction and classification characteristics of the deposit.

[0033] 3. Different posterior probability approximation methods have a great impact on the uncertainty of the prediction results. The present invention adopts the Monte Carlo Metropolis-Hastings sampling method to provide a more reliable uncertainty estimation, which can effectively distinguish cognitive uncertainty from accidental uncertainty. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present application will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 It is a Bayesian neural network architecture diagram in the embodiment of the present application; Figure 2 is the average probability map in the embodiment of the present application; Figure 3 is an epistemic uncertainty graph in an embodiment of the present application; Figure 4 is an epistemic uncertainty graph in an embodiment of the present application; Figure 5 is a performance evaluation diagram in an embodiment of the present application; Figure 6 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to have a clearer understanding of the technical features, purposes and effects of the present application, the specific implementation methods of the present application are now described in detail with reference to the accompanying drawings.

[0036] The embodiments of the present application provide a Bayesian deep learning method for mineral prediction and exploration risk assessment.

[0037] Please refer to Figure 1 , Figure 1 It is a Bayesian neural network architecture diagram of a Bayesian deep learning mineral prediction and exploration risk assessment method in an embodiment of the present application, including: S1: Construct a multi-source geological data set and preprocess it to obtain ore deposits and ore-controlling elements, and construct an ore-controlling element data set; S2: Construct a Bayesian neural network training model based on Metropolis-Hastings sampling; S3: Input the mining control factor data set into the Bayesian neural network training model to obtain the probability mean, aleatory uncertainty estimation, and epistemic uncertainty estimation; S4: Through the probability mean, aleatory uncertainty estimation and epistemic uncertainty estimation, the prediction probability map, aleatory uncertainty map and epistemic uncertainty map are drawn to evaluate the mineral exploration risk and prospecting potential.

[0038] In one embodiment of the present application, the classification performance of the Bayesian neural network training model is evaluated using ROC curves, AUC values, etc.

[0039]

[0040] Among them, TPR is the true positive rate and FPR is the false positive rate.

[0041] Step S1 includes: Multi-source geological data sets include: geological data, geochemical data, geographic data, and remote sensing data; Preprocessing includes: missing value processing, outlier processing and noise processing; The preprocessed multi-source address data set is screened to obtain mineral deposits and mineral-controlling elements.

[0042] Step S2 includes: Set the number of network layers, number of neurons in each layer, and activation function of the Bayesian neural network training model ; Set the prior distribution of weights and biases of the Bayesian neural network training model as follows: Taking the ore deposit and ore-controlling factors as input data, the input data of the Bayesian neural network training model is defined as , the output label is , the training data is ; The neural network parameters are The probability distribution of the predicted output of the Bayesian neural network training model is expressed as follows:

[0043]

[0044]

[0045] In the formula, Output predictions for the Bayesian neural network training model, is the activation function, They are the first The weight matrix and bias vector of the layer.

[0046] Step S2 also includes: Set the likelihood function of the Bayesian neural network training model. The specific steps are as follows: The presence of mineral deposits is marked as 1, and the absence of mineral deposits is marked as 0. Each sample belongs to one of the two categories; Assume that each sample Tags The value is 0 or 1, and the predicted probability output by the Bayesian neural network training model Representation sample The probability of belonging to category 1; The likelihood function for the binomial distribution is defined as:

[0047]

[0048] in, For sample The true label of express The probability of taking 0 is 0, The probability of taking 1 is ; The probability of taking 0 is , The probability of taking 1 is 0; Convert the likelihood function of the binomial distribution to the log-likelihood function:

[0049] Taking the negative log likelihood, we get the negative log likelihood, which is the loss function ,as follows: .

[0050] As an embodiment, the loss function is minimized to measure the difference between the probability distribution predicted by the Bayesian neural network training model and the true label distribution, and a loss value is obtained; the neural network parameters of the Bayesian neural network training model are adjusted according to the loss value.

[0051] Step S3 includes: S31: Setting the initial neural network parameters for the Bayesian neural network training model ; Select Proposal Distribution , specifically: using Gaussian distribution as a prior distribution;

[0052] in, is the covariance matrix, is a diagonal matrix; represents the Gaussian distribution function; represents the generated parameter samples; Indicates The parameters of the iterations; S32: Perform Metropolis-Hastings sampling to generate parameter samples; For each iterative sampling process, the generated parameter samples are as follows:

[0053]

[0054] in is the total number of iterative sampling processes; Calculating the acceptance rate ,as follows:

[0055] in, Indicated in the parameter Next, label data In the conditions The likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Indicates that the sampling parameters were obtained in the previous iteration Based on the label data In the conditions The likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Determine whether to accept the parameter sample. The specific steps are as follows: Generate a uniformly distributed random number ,like , then accept ,Right now:

[0056] like , then reject ,Right now:

[0057] S33: Approximate the posterior probability distribution by generating parameter samples :

[0058] in, represents training data; Indicates the given parameter Under these conditions, the observed data The likelihood function of ; Representation parameters The prior probability of S34: Save the parameter sample set obtained by sampling ; S35: Perform model prediction using a set of parameter samples, and calculate the probability mean of the prediction probability and the estimate of the aleatory uncertainty; S36: Calculate the variance of the predicted probability under different parameter samples to obtain the epistemic uncertainty estimate.

[0059] Step S35 includes: The steps to calculate the probability mean are as follows: For new input data , using all sampled parameter samples Perform forward propagation to obtain a series of predicted probabilities , the probability mean after M samplings for:

[0060] The calculation steps of the aleatory uncertainty estimate are as follows: By predicting the variance at the output layer of the Bayesian neural network training model , and obtain the aleatory uncertainty estimate ,as follows:

[0061] Among them, the variance Reflect input data The noise itself; Represents the input data The probability of getting Represents the input data The average probability obtained.

[0062] Step S36 includes: Calculate the variance of the predicted probability under different parameter samples:

[0063] in, represents the parameter sample of the tth sampling; Representing probability Average.

[0064] Step S4 includes: Draw a predicted probability map based on the probability mean to form a mineral potential distribution map; Create aleatoric uncertainty maps based on aleatoric uncertainty estimates to identify areas with high data noise; Draw epistemic uncertainty maps based on epistemic uncertainty estimates to identify areas where knowledge is insufficient or data is scarce; Combined with mineral potential distribution maps, areas with high data noise and areas with insufficient knowledge or scarce data, it assists in mineral exploration decision-making, determines priority exploration areas and areas that require further data collection, and completes mineral exploration risk and prospecting potential evaluation.

[0065] In one embodiment of the present application, based on a multi-source geological dataset in the Nanling region, taking tungsten polymetallic mineralization prediction and its uncertainty evaluation as an example, the implementation process of the technical method and its efficiency and practicality in practical applications are explained. The present invention is further explained below in conjunction with the accompanying drawings.

[0066] (1) Eleven ore-controlling elements in the Nanling region were selected as input data X, including northwest-trending faults, northeast-trending faults, east-west faults, north-south faults, fault density, granite, and five geochemical variables: W, Sn, Mo, Bi, and Ag.

[0067] (2) The coordinates of confirmed mineral deposits in the study area are collected as positive sample data, and a corresponding number of data points are randomly selected from areas without mineral deposits as negative sample data. The two together constitute the input label data Y.

[0068] (3) Based on GIS technology, the data of mineral control elements and label data are rasterized, with the grid unit of 1 km × 1 km, which is convenient for model processing and calculation; (4) Set the number of network layers, the number of neurons, and the activation function, and denote these learnable weights and biases as θ; give the prior distribution of θ , using the likelihood function of the binomial distribution , the Bayesian neural network structure constructed, its schematic diagram is as follows Figure 1 shown.

[0069] (5) Metropolis-Hastings is used for sampling, and the initialization parameters are , generating candidate parameters in the iterative process , calculate the acceptance rate , and decide whether to accept the sequence generated by Approximately follows the target posterior distribution.

[0070] (6) Use parameter sampling set to analyze new samples Do forward propagation to get the predicted probability distribution, take its mean, and you can get the predicted average probability distribution map of tungsten polymetallic deposits in the Nanling area, such as Figure 2 As shown; By calculating the variance, the epistemic uncertainty of the prediction results of tungsten polymetallic minerals is quantified, as shown in Figure 3 As shown; the output noise parameters can characterize the accidental uncertainty of the prediction results of tungsten polymetallic ore, such as Figure 4 shown.

[0071] (7) Use ROC curves, AUC values, etc. to evaluate the classification performance of the Bayesian neural network training model based on Metropolis-Hastings sampling, such as Figure 5 As shown, the model has excellent predictive ability.

[0072] The present application also discloses an electronic device. Figure 6 , Figure 6 The electronic device 500 may include: at least one processor 501 , at least one network interface 504 , a user interface 503 , a memory 505 , and at least one communication bus 502 .

[0073] The communication bus 502 is used to realize the connection and communication between these components.

[0074] The user interface 503 may include a display screen, and the optional user interface 503 may also include a standard wired interface or a wireless interface.

[0075] The network interface 504 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0076] The present application also discloses a computer-readable storage medium storing a plurality of instructions suitable for loading by a processor to execute the above-mentioned Bayesian deep learning method for mineral prediction and exploration risk assessment.

[0077] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure.

[0078] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A Bayesian deep learning method for mineral prediction and exploration risk assessment, characterized in that: The method comprises the following steps: S1: Construct a multi-source geological data set and preprocess it to obtain ore deposits and ore-controlling elements, and construct an ore-controlling element data set; S2: Construct a Bayesian neural network training model based on Metropolis-Hastings sampling; S3: Input the mining control factor data set into the Bayesian neural network training model to obtain the probability mean, aleatory uncertainty estimation, and epistemic uncertainty estimation; S4: Through the probability mean, aleatory uncertainty estimation and epistemic uncertainty estimation, the prediction probability map, aleatory uncertainty map and epistemic uncertainty map are drawn to evaluate the mineral exploration risk and prospecting potential.

2. The method for mineral prediction and exploration risk assessment based on Bayesian deep learning as claimed in claim 1, characterized in that: Step S1 includes: Multi-source geological data sets include: geological data, geochemical data, geographic data, and remote sensing data; Preprocessing includes: missing value processing, outlier processing and noise processing; The preprocessed multi-source address data set is screened to obtain mineral deposits and mineral-controlling elements.

3. The Bayesian deep learning method for mineral prediction and exploration risk assessment according to claim 1, characterized in that: Step S2 includes: Set the number of network layers, number of neurons in each layer, and activation function of the Bayesian neural network training model ; Set the prior distribution of weights and biases of the Bayesian neural network training model as follows: Taking the ore deposit and ore-controlling factors as input data, the input data of the Bayesian neural network training model is defined as , the output label is , the training data is ; The neural network parameters are The probability distribution of the predicted output of the Bayesian neural network training model is expressed as follows: In the formula, Output predictions for the Bayesian neural network training model, is the activation function, They are the first The weight matrix and bias vector of the layer.

4. The method for mineral prediction and exploration risk assessment based on Bayesian deep learning as claimed in claim 3, characterized in that: Step S2 also includes: Set the likelihood function of the Bayesian neural network training model. The specific steps are as follows: The presence of mineral deposits is marked as 1, and the absence of mineral deposits is marked as 0. Each sample belongs to one of the two categories; Assume that each sample Tags The value is 0 or 1, and the predicted probability output by the Bayesian neural network training model Representation sample The probability of belonging to category 1; The likelihood function for the binomial distribution is defined as: in, For sample The true label of express The probability of taking 0 is 0, The probability of taking 1 is ; The probability of taking 0 is , The probability of taking 1 is 0; Convert the likelihood function of the binomial distribution to the log-likelihood function: Taking the negative log likelihood, we get the negative log likelihood, which is the loss function ,as follows: 。 5. The method for mineral prediction and exploration risk assessment based on Bayesian deep learning as claimed in claim 4, characterized in that: Step S3 includes: S31: Setting the initial neural network parameters for the Bayesian neural network training model ; Select Proposal Distribution , specifically: using Gaussian distribution as a prior distribution; in, is the covariance matrix, is a diagonal matrix; represents the Gaussian distribution function; represents the generated parameter samples; Indicates The parameters of the iterations; S32: Perform Metropolis-Hastings sampling to generate parameter samples; For each iterative sampling process, the generated parameter samples are as follows: in is the total number of iterative sampling processes; Calculating the acceptance rate ,as follows: in, Indicated in the parameter Next, label data In the conditions The likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Indicates that the sampling parameters were obtained in the previous iteration Based on the label data In the conditions The likelihood function under ; Representation parameters The prior probability distribution of ; Represents the distribution from proposals In Generate for condition The probability density of Determine whether to accept the parameter sample. The specific steps are as follows: Generate a uniformly distributed random number ,like , then accept ,Right now: like , then reject ,Right now: S33: Approximate the posterior probability distribution by generating parameter samples : in, represents training data; Indicates the given parameter Under these conditions, the observed data The likelihood function of ; Representation parameters The prior probability of S34: Save the parameter sample set obtained by sampling ; S35: Perform model prediction using a set of parameter samples, and calculate the probability mean of the prediction probability and the estimate of the aleatory uncertainty; S36: Calculate the variance of the predicted probability under different parameter samples to obtain the epistemic uncertainty estimate.

6. A Bayesian deep learning method for mineral prediction and exploration risk assessment as claimed in claim 5, characterized in that: Step S35 includes: The steps to calculate the probability mean are as follows: For new input data , using all sampled parameter samples Perform forward propagation to obtain a series of predicted probabilities , the probability mean after M samplings for: The calculation steps of the aleatory uncertainty estimate are as follows: By predicting the variance at the output layer of the Bayesian neural network training model , and obtain the aleatory uncertainty estimate ,as follows: Among them, the variance Reflect input data The noise itself; Represents the input data The probability of getting Represents the input data The average probability obtained.

7. The Bayesian deep learning method for mineral prediction and exploration risk assessment according to claim 5, characterized in that: Step S36 includes: Calculate the variance of the predicted probability under different parameter samples: in, represents the parameter sample of the tth sampling; Representing probability Average.

8. The Bayesian deep learning method for mineral prediction and exploration risk assessment according to claim 1, characterized in that: Step S4 includes: Draw a predicted probability map based on the probability mean to form a mineral potential distribution map; Create aleatoric uncertainty maps based on aleatoric uncertainty estimates to identify areas with high data noise; Draw epistemic uncertainty maps based on epistemic uncertainty estimates to identify areas where knowledge is insufficient or data is scarce; Combined with mineral potential distribution maps, areas with high data noise and areas with insufficient knowledge or scarce data, it assists in mineral exploration decision-making, determines priority exploration areas and areas that require further data collection, and completes mineral exploration risk and prospecting potential evaluation.

9. An electronic device, characterized in that: It includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed by a computer, the method according to any one of claims 1 to 8 is executed.

Images