Tooth extraction postoperative complication prediction method and system based on DBSCAN-BP neural network
By combining DBSCAN-BP neural networks, the problems of difficulty in learning nonlinear relationships and lack of consideration of mutual constraints in the prediction of post-extraction complications are solved, achieving high-precision complication prediction, guiding medical staff to take effective nursing measures, and improving patient recovery efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2023-09-28
- Publication Date
- 2026-07-14
AI Technical Summary
Existing methods for predicting post-extraction complications rely on domain expert knowledge, struggle to learn nonlinear relationships between features, have low prediction accuracy, and do not consider the interrelationships between different complications.
By combining DBSCAN and BP neural networks, the DBSCAN algorithm is used to cluster physiological feature data, and the BP neural network is used for hidden layer deep learning. The mutual constraints of feature data are considered to improve the convergence speed and accuracy of the prediction model.
It improves the accuracy and reliability of predicting post-extraction complications, guides medical staff to take targeted measures, avoids serious complications, and improves patient recovery efficiency and treatment outcomes.
Smart Images
Figure CN117352160B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of medical technology, specifically to a method and system for predicting post-extraction complications based on a DBSCAN-BP neural network. Background Technology
[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.
[0003] With the development of science and technology, artificial intelligence technology, represented by deep neural networks, has developed rapidly. After model construction and training, it can effectively extract knowledge from data and predict spatial and spatiotemporal correlations. Using deep neural networks trained on large amounts of data, it can even predict and diagnose disease complications. As people's living standards have significantly improved, their demands for oral aesthetics and hygiene have increased, promoting the development of dentistry. Tooth extraction is a widely used procedure in oral surgery. Due to the complex local anatomy of the oral cavity and limited operating space, the incidence of complications after tooth extraction in outpatient dental surgery is extremely high. Some complications can be eliminated through active and effective postoperative care, but some are more difficult to manage. Common postoperative complications after tooth extraction include postoperative bleeding, swelling, and pain, which can also lead to fluctuations in blood pressure, body temperature, and pulse. Patients with cardiovascular disease or blood disorders should be especially careful, otherwise, serious consequences may result. Tooth loss can cause alveolar bone atrophy, displacement or elongation of adjacent teeth, causing chewing difficulties. Loss of anterior teeth directly affects pronunciation and appearance, etc. Failure to treat these complications in a timely manner can seriously affect the patient's treatment outcome and physical and mental health. Meanwhile, medical staff need to take targeted diagnostic and treatment measures and nursing methods for different complications, so as to maximize the recovery efficiency of patients after tooth extraction and ensure the treatment effect.
[0004] Numerous research achievements have been made in the field of postoperative risk prediction, such as the invention patent CN108717869B - Diagnostic Auxiliary System for Diabetic Retinal Complications Based on Convolutional Neural Network; the invention patent CN 114242234B - Prediction Method for Postoperative Complication Risk Value of TAVR Based on Aggregation Neural Network; and the invention patent CN 116434968A - A Multi-Label Fine-Grained Postoperative Complication Prediction Device with Hierarchical Clustering Constraints, all of which have improved the accuracy of postoperative risk prediction.
[0005] However, the inventors discovered that most current methods for predicting postoperative risk belong to traditional statistical machine learning methods, which face several problems:
[0006] First, it requires relying on the knowledge of domain experts for feature selection and labeling;
[0007] Second, traditional machine learning is mostly linear models, making it difficult to learn non-linear relationships between features. In addition, existing neural network-based risk predictions mostly predict the general situation of risk occurrence, such as high risk or low risk, but cannot derive detailed probability values of risk occurrence. The accuracy of risk prediction is low, and the conclusions of the predictions have limited reference value. Moreover, current research models do not take into account the mutual constraints between different complications. For example, the occurrence of some complications may reduce the probability of other complications, while the occurrence of some complications may increase the probability of other complications. Summary of the Invention
[0008] To address the aforementioned issues, this disclosure proposes a method and system for predicting post-extraction complications based on a DBSCAN-BP neural network. The method uses cluster analysis feature data as input to the BP neural network, employs deep learning in hidden layers, and outputs post-extraction complications. The DBSCAN algorithm and BP neural network are combined in the prediction model, taking into account the mutual constraints of the feature data. This improves the convergence speed and effectiveness of the prediction model, thereby enhancing the accuracy and reliability of the prediction.
[0009] According to some embodiments, the present disclosure adopts the following technical solutions:
[0010] A method for predicting post-extraction complications based on a DBSCAN-BP neural network includes:
[0011] Obtain postoperative physiological characteristic data of patients and perform feature clustering on the obtained physiological characteristic data;
[0012] The physiological characteristic data of each feature data cluster is normalized and optimized.
[0013] A static BP neural network model is constructed, with the feature data cluster as input. By adjusting the eps and MinPts parameters in the clustering algorithm, the feature data cluster is automatically updated, and the parameters of the static BP neural network model are modified to output the evaluation prediction results.
[0014] According to some embodiments, the present disclosure adopts the following technical solutions:
[0015] A DBSCAN-BP neural network-based system for predicting post-extraction complications includes:
[0016] The data acquisition module is used to acquire postoperative physiological characteristic data of patients and perform feature clustering on the acquired physiological characteristic data;
[0017] The DBSCAN clustering module is used to normalize and optimize the physiological feature data of each feature data cluster.
[0018] The evaluation and prediction module is used to construct a static BP neural network model. It takes the feature data clusters as input to the static BP neural network model, automatically updates the feature data clusters by adjusting the eps and MinPts parameters in the clustering algorithm, modifies the parameters of the static BP neural network model, and outputs the evaluation and prediction results.
[0019] According to some embodiments, the present disclosure adopts the following technical solutions:
[0020] A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the aforementioned method for predicting post-extraction complications based on a DBSCAN-BP neural network.
[0021] According to some embodiments, the present disclosure adopts the following technical solutions:
[0022] An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the method for predicting post-extraction complications based on a DBSCAN-BP neural network.
[0023] Compared with the prior art, the beneficial effects of this disclosure are as follows:
[0024] This disclosure presents a method for predicting post-extraction complications based on a DBSCAN-BP neural network. It employs the DBSCAN clustering algorithm, defining a cluster as the largest set of densely connected points. This algorithm can divide regions with sufficiently high density into clusters and can discover clusters of arbitrary shapes in noisy spatial databases. It collects a large amount of patient physiological characteristic data and obtains feature data clusters based on the clustering algorithm. The BP neural network consists of forward and backward propagation processes. In the forward propagation process, the input pattern is processed layer by layer from the input layer through the hidden unit layers and then forward to the output layer. The state of each neuron in each layer only affects the state of the next layer. If the desired output cannot be obtained in the output layer, backward propagation is initiated. Backward propagation involves transmitting the error signal back along the original forward propagation path and modifying the weight coefficients of each neuron in each hidden layer to minimize the error signal. Combining the DBSCAN algorithm with the BP neural network leverages their respective advantages. Cluster analysis considers the interrelationships between different complications; for example, the occurrence of some complications may reduce the probability of other complications, while the occurrence of some complications may increase the probability of other complications. Backpropagation (BP) neural networks possess the ability to approximate arbitrary nonlinear mappings through learning. Applying neural networks to the identification and prediction of nonlinear systems is not limited by nonlinear models, facilitating the development of easily implemented learning algorithms for assessing, predicting, and diagnosing medical complications. The DBSCAN-BP neural network static model considers the mutual constraints of feature data, improving the convergence speed and effectiveness of the prediction model, enhancing its accuracy and reliability, and better guiding medical staff to take targeted measures. This helps to minimize serious post-extraction complications and avoid severely impacting patients' treatment outcomes and mental and physical health. Furthermore, while medical staff may provide comprehensive instructions to patients potentially experiencing multiple complications, patients with varying memory and comprehension abilities may not fully accept all instructions, leading to a poor medical experience. Predictive diagnosis can guide medical staff to provide targeted treatment and nursing recommendations, maximizing post-extraction recovery efficiency and ensuring treatment effectiveness. Attached Figure Description
[0025] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.
[0026] Figure 1 This is a flowchart of a method for predicting post-extraction complications based on a DBSCAN-BP neural network according to an embodiment of the present disclosure.
[0027] Figure 2This is a structural diagram of a tooth extraction postoperative complication prediction system based on a DBSCAN-BP neural network according to an embodiment of this disclosure. Detailed Implementation
[0028] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.
[0029] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.
[0030] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0031] Example 1
[0032] One embodiment of this disclosure provides a method for predicting post-extraction complications based on a DBSCAN-BP neural network, including:
[0033] Step 1: Obtain postoperative physiological characteristic data of patients and perform feature clustering on the obtained physiological characteristic data;
[0034] Step 2: Normalize and optimize the physiological feature data of each feature data cluster;
[0035] Step 3: Construct a static BP neural network model. Use the feature data clusters as input to the static BP neural network model. By adjusting the eps and MinPts parameters in the clustering algorithm, the feature data clusters are automatically updated, and the parameters of the static BP neural network model are modified. The evaluation prediction results are then output.
[0036] As one embodiment, a specific implementation method for predicting post-extraction complications based on a DBSCAN-BP neural network includes:
[0037] Step 1: Obtain postoperative physiological characteristic data of patients and perform feature clustering on the obtained physiological characteristic data;
[0038] Specifically, this includes collecting a large amount of physiological data on the patient's facial and oral structure (whether there is redness and swelling in the cheeks, swelling of lymph nodes under the jaw and in the neck, numbness or other abnormal sensations in the lower lip, etc.), intraoral examination results (tooth growth, whether there is inflammation around the teeth, whether there is decay in the surrounding teeth, etc.), physical examination results (height, weight, etc.), biochemical test results (blood pressure, blood lipids, blood sugar, etc.), and various postoperative complications (bleeding, swelling, pain, fluctuations in blood pressure, body temperature, and pulse affecting patients with cardiovascular and blood diseases, tooth loss can cause alveolar bone atrophy, displacement or elongation of adjacent and opposing teeth, causing chewing difficulties, and anterior tooth loss directly affecting pronunciation and appearance, etc.).
[0039] Step 2: Normalize and optimize the physiological feature data of each feature data cluster;
[0040] Furthermore, DBSCAN is a representative density-based clustering algorithm. Unlike partitioning and hierarchical clustering methods, it defines a cluster as the largest set of density-connected points. It can divide regions with sufficiently high density into clusters and can discover clusters of arbitrary shapes in noisy spatial databases. The DBSCAN algorithm has relatively simple steps: the first step is to scan all sample points to find core points and form temporary clusters; the second step is to merge temporary clusters to obtain full clusters, until every point in the current temporary cluster is either not in the core point list or all its directly reachable points are already in that temporary cluster, at which point the temporary cluster is upgraded to a full cluster. This algorithm has the following advantages: first, it does not require prior knowledge of the number of clusters to be formed; second, it can discover clusters of arbitrary shapes; third, it can identify noise points and is robust to noise points far from the density core; and fourth, it is insensitive to the order of samples in the database.
[0041] Furthermore, the specific steps of using DBSCAN to perform feature clustering on the acquired physiological feature data in this disclosure are as follows:
[0042] Obtain as comprehensive a range of physiological characteristics as possible from the patient, including but not limited to: whether there is redness and swelling in the cheeks, whether there is swelling of the lymph nodes under the jaw and in the neck, whether there is numbness or other abnormal sensations in the lower lip, the growth of teeth, whether there is inflammation around the teeth, whether there is decay in the surrounding teeth, height, weight, blood pressure, blood lipids, blood sugar, etc. Classify the physiological characteristics according to oral structure, intraoral examination results, physical examination results, and biochemical test results, and cluster the characteristics according to the inherent relationships of the characteristic data.
[0043] Furthermore, the specific implementation process for normalizing and optimizing the physiological characteristic data of each feature data cluster is as follows:
[0044] All feature data have their own dimensions. Normalization aims to eliminate these dimensions, enabling comparability between features with different dimensions and mitigating the impact of dimensional differences on the analysis results. Unnormalized feature values that are too large will cause computational problems. Normalization maps all feature data with different dimensions to a unified [0, 1] linear relationship. This linear transformation does not change the original properties of the data, reducing the computational load of the next step in the BP neural network, increasing the convergence speed, and improving the convergence accuracy. In short, normalization maps feature data with different dimensions to a unified standard without altering the original data's properties, normalizing all feature values to a single standard, thus improving the convergence speed and accuracy of the neural network.
[0045] Step 3: Construct a static model of the BP neural network, including: A typical BP neural network consists of three feedforward layers: an input layer, intermediate layers (also called hidden layers), and an output layer. Neurons in each layer are fully connected only to neurons in adjacent layers; there are no connections between neurons within the same layer, and no feedback connections between neurons in different layers, forming a hierarchical feedforward neural network system. The computation process of a BP neural network consists of forward propagation and backward propagation. In forward propagation, the input pattern is processed layer by layer from the input layer through the hidden unit layers and then forward to the output layer. The state of each neuron in one layer only affects the state of the neuron in the next layer. If the desired output cannot be obtained in the output layer, backward propagation begins. Backward propagation involves transmitting the error signal back along the original forward propagation path and modifying the weights of each neuron in each hidden layer to minimize the error signal.
[0046] Furthermore, the process of training the static model of the BP neural network includes:
[0047] (1) Collect a large amount of patient characteristic information, including but not limited to facial and oral structure (whether there is redness and swelling of the cheeks, whether there is swelling of the lymph nodes under the jaw and neck, whether there is numbness or other abnormal sensations in the lower lip, etc.), intraoral examination results (the growth of teeth, whether there is inflammation around the teeth, whether there is decay of the surrounding teeth, etc.), physical examination results (height, weight, etc.), biochemical test results (blood pressure, blood lipids, blood sugar, etc.), and various postoperative complications (bleeding, swelling, pain, fluctuations in blood pressure, body temperature, and pulse that affect patients with cardiovascular disease and blood diseases, tooth loss can cause alveolar bone atrophy, displacement or elongation of adjacent teeth and opposing teeth, causing chewing difficulties, and the loss of anterior teeth directly affects pronunciation and appearance, etc.).
[0048] (2) Specify the eps and MinPts algorithm parameters, and use the DBSCAN clustering algorithm to obtain multiple feature data clusters;
[0049] The DBSCAN clustering algorithm uses two feature values: eps and MinPts. eps represents the center of the cluster, and MinPts represents the radius of the cluster. Specifying eps and MinPts determines the amount of data contained in the cluster. Since different clusters contain different feature data, different eps and MinPts are set to define each cluster.
[0050] (3) Normalize the feature data of each data cluster class to further optimize the feature data cluster class;
[0051] Normalization of data cluster feature data removes dimensions and establishes a unified mapping relationship between data in each cluster without changing the nature of the data. This further optimizes the feature data of the data clusters to facilitate the improvement of accuracy and speed in the next step of the BP neural network.
[0052] (4) Divide the multiple feature data clusters described in step (3) into a training sample set and a test sample set;
[0053] (5) The normalized facial and oral structure, intraoral examination results, physical examination results, biochemical examination results and other feature data clusters are used as inputs, and postoperative complications are used as outputs to establish a static model. The number of hidden layers and the number of neurons are set. The BP neural network is trained using the training sample set to obtain the BP neural network static model.
[0054] (6) By adjusting the eps and MinPts algorithm parameters of the DBSCAN clustering algorithm in step (2), the data clusters are automatically updated, thereby improving the convergence speed of the BP neural network.
[0055] By changing eps and MinPts to alter the number of values in each cluster, the size of the cluster circle is updated through feedback from the BP neural network, which represents the number of values in each cluster. This achieves the goal of updating the data cluster categories and improves the convergence speed of the BP neural network.
[0056] (7) Use the test sample set from step (4) to input the BP neural network model to test the accuracy of the model, modify the parameters of the BP neural network, and improve the convergence accuracy of the BP neural network.
[0057] A backpropagation (BP) neural network consists of an input layer, a hidden layer, and an output layer. The weights of each neuron in the hidden layer are automatically generated, and the accuracy of the BP neural network can be improved by changing the weight values of each neuron.
[0058] (8) Repeat steps (6)-(7) until the iteration termination condition is met and the training reaches the model accuracy requirement.
[0059] The termination condition is to achieve the required accuracy by changing the eps and MinPts of the DBSCAN clustering algorithm and the weight values of each neuron in the hidden layer of the BP neural network.
[0060] (9) Output the evaluation and prediction results of the DBSCAN-BP neural network. The prediction results are the probability values of the form and severity of complications that the patient may experience through clinical treatment. Based on the prediction results, targeted medical and nursing suggestions are made for the patient to avoid serious complications after tooth extraction as much as possible and to avoid seriously affecting the patient's treatment effect and physical and mental health.
[0061] Example 2
[0062] One embodiment of this disclosure provides a prediction system for post-extraction complications based on a DBSCAN-BP neural network, comprising:
[0063] The data acquisition module is used to acquire postoperative physiological characteristic data of patients and perform feature clustering on the acquired physiological characteristic data;
[0064] The DBSCAN clustering module is used to normalize and optimize the physiological feature data of each feature data cluster.
[0065] The evaluation and prediction module is used to construct a static BP neural network model. It takes the feature data clusters as input to the static BP neural network model, automatically updates the feature data clusters by adjusting the eps and MinPts parameters in the clustering algorithm, modifies the parameters of the static BP neural network model, and outputs the evaluation and prediction results.
[0066] Furthermore, it also includes a sample set creation module, which is used to divide the feature data clusters into training sample sets and test sample sets;
[0067] A BP network training module is used to input the training sample set into the BP network for network training.
[0068] The DBSCAN-BP neural network optimization module is used to input the test sample set into the BP neural network and improve the convergence speed and accuracy of the prediction model by adjusting the eps and MinPts algorithm parameters.
[0069] Example 3
[0070] One embodiment of this disclosure provides a non-transitory computer-readable storage medium for storing computer instructions. When these computer instructions are executed by a processor, they implement the method for predicting post-extraction complications based on a DBSCAN-BP neural network.
[0071] Example 4
[0072] One embodiment of this disclosure provides an electronic device, including a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the method for predicting post-extraction complications based on the DBSCAN-BP neural network.
[0073] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0074] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0075] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.
Claims
1. A method for predicting post-extraction complications based on a DBSCAN-BP neural network, characterized in that, include: Obtain postoperative physiological characteristic data of patients and perform feature clustering on the obtained physiological characteristic data; The physiological characteristic data of each feature data cluster is normalized and optimized. A static BP neural network model is constructed, and the feature data clusters are used as input to the static BP neural network model. By adjusting the eps and MinPts parameters in the clustering algorithm, the feature data clusters are automatically updated, and the parameters of the static BP neural network model are modified to output the evaluation prediction results. The size of the cluster circle is updated by using backpropagation (BP) neural network feedback, which improves the convergence speed of the BP neural network.
2. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 1, characterized in that, The physiological characteristic data includes facial and oral cavity structural characteristics, intraoral examination results, physical examination results, biochemical test results, and characteristic data of factors affecting various postoperative complications.
3. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 1, characterized in that, The DBSCAN algorithm is used to perform feature clustering on physiological feature data. All physiological feature data are scanned to find core points to form temporary clusters. Temporary clusters are merged to obtain a cluster. When every point in the current temporary cluster is no longer in the list of core points, or when all points directly reachable by its density are already in the temporary cluster, the temporary cluster is upgraded to a cluster.
4. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 1, characterized in that, The constructed BP neural network static model includes three layers of feedforward network: input layer, intermediate layer and output layer. The neurons in each layer are fully connected only to the neurons in the adjacent layers. There are no connections between neurons in the same layer and no feedback connections between neurons in different layers, thus forming a hierarchical feedforward neural network static model.
5. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 4, characterized in that, The static model of a BP neural network consists of forward and backward computation processes. In the forward propagation process, the input is processed layer by layer from the input layer through the hidden unit layer and then forward to the output layer. The state of each neuron only affects the state of the next neuron. If the expected output cannot be obtained in the output layer, the backward propagation process begins.
6. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 5, characterized in that, Backpropagation involves transmitting the error signal back along the original forward propagation path and modifying the weight coefficients of each neuron in each hidden layer to minimize the error signal.
7. The method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in claim 1, characterized in that, The training process of a BP neural network static model includes: dividing the feature data cluster into a training sample set and a test sample set; using the feature data cluster as input; setting the number of hidden layers and the number of neurons in each layer; training the BP neural network using the training sample set to obtain the BP neural network static model; using the test sample set as input to the BP neural network static model to test the model accuracy; modifying the BP neural network parameters to improve convergence accuracy; and continuing until the iteration termination condition is met to achieve the required model accuracy.
8. A system for predicting post-extraction complications based on a DBSCAN-BP neural network, characterized in that, include: The data acquisition module is used to acquire postoperative physiological characteristic data of patients and perform feature clustering on the acquired physiological characteristic data; The DBSCAN clustering module is used to normalize and optimize the physiological feature data of each feature data cluster. The evaluation and prediction module is used to construct a static BP neural network model. It takes the feature data clusters as input to the static BP neural network model, automatically updates the feature data clusters by adjusting the eps and MinPts parameters in the clustering algorithm, modifies the parameters of the static BP neural network model, and outputs the evaluation and prediction results.
9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in any one of claims 1-7.
10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the method for predicting post-extraction complications based on a DBSCAN-BP neural network as described in any one of claims 1-7.