Decision tree growth method, system, and device based on association rules

Abstract

This invention belongs to the field of command and control system construction decision-making, specifically relating to a method, system, and device for decision tree growth based on association rules. The method includes: acquiring original data samples; extracting mining features based on the original data samples; discretizing the data in the original data samples based on the mining features to obtain transaction items corresponding to different mining features, with different transaction items constituting mining transactions; performing association rule mining on the transactions to be mined to obtain association rules between transaction items; obtaining an initial decision tree based on the mining features; and integrating the association rules into the initial decision tree to obtain an optimized decision tree. This application has the effect of reducing interference from invalid information and rapidly constructing decision trees.

Description

Technical Field

[0001] This invention belongs to the field of command and control system construction decision-making, specifically involving a decision tree growth method, system, and device based on association rules. Background Technology

[0002] In future multi-domain and intelligent combat scenarios, the resources available to command and control systems are characterized by wide distribution, diverse sources, multiple types, complex relationships, and cross-domain nature. This significant expansion in types and spatiotemporal scope leads to an exponential increase in the scale and complexity of system construction. When faced with changes in combat formations, command relationships, and command targets, the current average reconfiguration time for command and control systems ranges from several hours to several days, making it difficult to adapt to the rapid pace of future operations. With the development of software-defined technology and the deepening of information technology construction in the Chinese military, the proportion of digital and information-based components in equipment is increasing, and application functions are constantly expanding. Software-defined command and control systems have become one of the important future development directions for command information system architecture. Software-defined command and control systems require the system to autonomously perceive environmental changes, autonomously adjust its architecture, achieve human-environment-based combat control, and continuously acquire and learn dynamic knowledge to ensure the system's adaptability to continuously grow in combat capabilities. During the operation of a software-defined command and control system, changes in the operational environment, such as adjustments to combat missions, system node failures, or degradation of network communication environments, may cause the original organizational structure, business processes, task allocation methods, and command and control links of the command and control system to become unsuitable for the new operational environment. Therefore, the system needs to make adaptive adjustment decisions autonomously and in a timely manner, generate adjustment plans, and implement structural or service adjustments.

[0003] In related technologies, adaptive decision-making in command and control systems traditionally employs policy-based or model-based decision-making methods. Policy-based decision-making refers to making decisions based on pre-defined policies. A policy describes the actions to be taken under given circumstances and conditions. The decision-making process is based on the selection and execution of policies to achieve the desired goal. Model-based decision-making refers to making decisions based on established mathematical or logical models. These models can be statistical models, mathematical models, or simulation models. By analyzing and predicting the models, the optimal decision-making solution can be derived.

[0004] Regarding the aforementioned technologies, both policy-based and model-based decision-making methods contain a large amount of invalid information during the decision-making process, which can affect the construction and performance of decision trees. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method, system and device for decision tree growth based on association rules. By mining association rules to construct decision trees, the interference of invalid information or information with weak correlation that is not relevant to radar blind spot filling is reduced, and the prediction results obtained by the decision tree are more accurate.

[0006] A decision tree growth method based on association rules, comprising:

[0007] Obtain raw data samples;

[0008] Based on the original data sample, extract and mine features;

[0009] Based on the mining features, the data in the original data sample is discretized to obtain transaction items corresponding to different mining features, and different transaction items constitute transactions to be mined;

[0010] Perform association rule mining on the transactions to be mined to obtain the association rules between transaction items;

[0011] Obtain the initial decision tree for the mined features;

[0012] The association rules are incorporated into the initial decision tree to obtain the optimized decision tree.

[0013] The original data samples are data related to insufficient radar detection area coverage, including data sequence number, number of blind spots, alarm level data, and blind spot area data. The features mined are the features of the samples in the original samples, including the number of blind spots, blind spot area, and alarm level.

[0014] Association rules are used to uncover relationships between features, representing the probability that one feature also exists given another. For example, given a certain number of blind zones, a certain blind zone area may also occur; given a certain blind zone area, a certain alarm level may also occur, and so on. A transaction item is a set of data items mined through association rules. The initial decision tree uses the number of blind zones, blind zone area, and alarm level as opportunity nodes, and specific solutions as strategy selection nodes. The solution determines the number and area of ​​blind spots to be filled for each radar.

[0015] Specifically, in decision theory, a decision tree is a predictive model that represents a mapping between object attributes and object values. Each node in the tree represents an object, each branch path represents a possible attribute value, and each leaf node corresponds to the value of the object represented by the path from the root node to that leaf node.

[0016] A decision tree is a hierarchical tree structure consisting of a root node, branches, internal nodes, and leaf nodes. It contains three types of nodes: alternative decision nodes, opportunity nodes, and strategy selection nodes.

[0017] In this context, the root node, or solution decision node, represents a branch that makes decisions based on certain problem characteristics and a specific problem decomposition method, thus representing a solution decision for the problem. For example, for the problem of insufficient radar detection area coverage, the solution decision node could be which combinations all areas should be divided into, and which areas should be included in each combination.

[0018] Internal nodes, also known as opportunity decision nodes, judge the decision-making process based on the data attributes related to the decision problem, and represent the path of the decision-making process. For example, for the problem of insufficient radar detection area coverage, opportunity decision nodes could be the radar blind zone area, the number of radar blind zones, the number of available radars, etc.

[0019] Leaf nodes, or policy selection nodes, can determine policies based on certain problem characteristics. The leaf nodes of a decision tree represent meta-policies rather than the final implementation policies.

[0020] Decision tree growth refers to the process of constructing a decision tree by recursively partitioning a dataset. In special circumstances, it can quickly reconstruct the system to generate decision strategies and achieve autonomous and adaptive decision-making capabilities.

[0021] Specifically, a decision tree growth method based on association rules also includes:

[0022] Obtain raw data samples of insufficient radar detection area coverage. The raw data samples include data sequence number, blind zone quantity data, alarm level data, and blind zone area data.

[0023] Based on the original data sample, mining features are extracted, including the number of blind spots, the area of ​​blind spots, and the alarm level.

[0024] Based on the mining features, the data in the original data sample is discretized to obtain transaction items corresponding to different mining features, and different transaction items constitute transactions to be mined;

[0025] The association rule mining is performed on the transaction to be mined to obtain the association rule between the transaction items. The association rule is the probability of the other mining features appearing when one mining feature is used as a condition.

[0026] Obtain an initial decision tree for the mined features, the initial decision tree including the hierarchical order of the mined features in the decision tree; integrate the association rules into the initial decision tree to obtain an optimized decision tree.

[0027] Optionally, the step of discretizing the data in the original data sample based on the mining features to obtain transaction items corresponding to different mining features includes:

[0028] Initialize the K cluster centers of the original data sample;

[0029] Calculate the spatial distance from each data point in the original data sample to the K cluster centers, and assign each data point to the cluster containing the nearest cluster center;

[0030] Calculate the average distance from all data points in a cluster to the cluster center, and use the average distance as the new cluster center;

[0031] Determine whether the new cluster center is the same as the original cluster center;

[0032] If the new cluster center is the same as the original cluster center, then the clustering result for each of the mining features is obtained;

[0033] Based on the original data sample, the contour coefficient is calculated;

[0034] The optimal contour coefficient is obtained by changing the K value based on the contour coefficient.

[0035] Obtain the optimal clustering result corresponding to the optimal contour coefficient, and use the optimal clustering result as the transaction item corresponding to different mining features.

[0036] Optionally, the step of changing the K value based on the contour coefficient to obtain the optimal contour coefficient includes:

[0037] Calculate the profile coefficient for different K values;

[0038] The contour coefficients under different K values ​​are compared to obtain the maximum contour coefficient, which is then taken as the optimal contour coefficient.

[0039] Optionally, the step of performing association rule mining on the transactions to be mined to obtain association rules between transaction items includes:

[0040] Calculate the support level for each transaction item;

[0041] Scan the transactions to be mined to obtain a candidate set;

[0042] Based on the candidate itemsets and the support, frequent itemsets are obtained;

[0043] Calculate the degree of interest among different transaction items in the frequent itemset to obtain the association rules among the transaction items.

[0044] Optionally, incorporating the association rules into the initial decision tree to obtain the optimized decision tree includes:

[0045] Obtain the mining feature level of the initial decision tree;

[0046] Obtain the mining feature level and value range corresponding to the mining feature of the association rule;

[0047] Based on the mining feature level and value range, the association rule is integrated into the initial decision tree to obtain an optimized decision tree.

[0048] Optionally, the step of integrating the association rules into the initial decision tree based on the mined feature level and value range to obtain the optimized decision tree includes:

[0049] The mined feature level is matched with the mined feature level to obtain the merged node of the mined level;

[0050] Based on the integration node and the value range, the association rule is integrated into the initial decision tree to obtain the optimized decision tree.

[0051] Optionally, the step of discretizing the data in the original data sample based on the mined features includes:

[0052] Obtain the data types corresponding to the different mined features;

[0053] Determine whether the data type is discrete data;

[0054] If the data type is discrete data, then the data in the original data sample is discretized based on the mining features.

[0055] A decision tree growth system based on association rules, comprising:

[0056] The acquisition module is used to acquire raw data samples;

[0057] The extraction module is used to extract and mine features based on the original data sample;

[0058] The discrete module is used to discretize the data in the original data sample based on the mining features to obtain transaction items corresponding to different mining features, and the different transaction items constitute the transaction to be mined;

[0059] The mining module is used to perform association rule mining on the transaction to be mined, and obtain the association rules between transaction items;

[0060] The second acquisition module is used to acquire the initial decision tree of the mined features;

[0061] A construction module is used to integrate the association rules into the initial decision tree to obtain an optimized decision tree.

[0062] A terminal device includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it employs a decision tree growth method based on association rules.

[0063] A computer-readable storage medium storing a computer program that, when loaded and executed by a processor, employs a decision tree growth method based on association rules.

[0064] The beneficial effects of this invention are as follows: By mining the original data samples, i.e., the data related to insufficient radar detection area coverage, mining features are obtained, namely the number of blind spots, the area of ​​blind spots, and the alarm level. Then, the continuous data in the mining features are discretized to obtain transaction items. Then, association rule mining is performed on the transactions to be mined, and the mined association rules are merged with the initial decision tree to obtain an optimized decision tree. Based on the initial decision tree, it is possible to predict the probability of the occurrence of another mining feature when one feature is a condition, thus realizing the growth of the decision tree. This provides a solution to the problem of insufficient radar detection area coverage, and allocates the number of blind spots to be filled and the blind spot area for each radar. By transforming the decision tree growth problem into an association rule mining and analysis problem, the mining and growth of association rules for system-adaptive decision trees in a big data environment is realized. Compared with traditional system-adaptive decision trees, since the mined association rules are all related, the interference of invalid information can be reduced, making the prediction results of the decision tree more accurate. Attached Figure Description

[0065] Figure 1 This is a flowchart illustrating a decision tree growth method based on association rules according to the present invention.

[0066] Figure 2 This is a schematic diagram illustrating the discretization of the original data samples in this invention;

[0067] Figure 3 This is a schematic diagram of the autonomous adaptive strategy decision tree of the present invention;

[0068] Figure 4 This is a schematic diagram of the initial decision tree of the present invention;

[0069] Figure 5 This is a schematic diagram of the optimized decision tree of the present invention. Detailed Implementation

[0070] A decision tree growth method based on association rules, such as Figure 1 As shown, it includes:

[0071] S100, Obtain raw data samples.

[0072] Specifically, the raw data sample includes data sequence numbers and different mining features, usually presented in tabular form. In this embodiment, taking the problem of insufficient radar detection area coverage as an example, the raw data sample includes the number of blind spots, the area of ​​the blind spots, and the alarm level, usually displayed in tabular form, such as... Figure 2 As shown.

[0073] S110. Extract and mine features based on the original data samples.

[0074] Specifically, the features to be mined are the characteristics or attributes of the sample, such as the number of blind spots, the area of ​​blind spots, and the alarm level. The number of blind spots, the area of ​​blind spots, and the alarm level are the features to be mined.

[0075] S120. Based on the mining features, the data in the original data sample is discretized to obtain the transaction items corresponding to different mining features. The different transaction items constitute the transaction to be mined.

[0076] Specifically, different mining features correspond to different data types. The number of blind spots is a discrete value, ranging from [1, 9]; the total area of ​​blind spots is a continuous value, ranging from (0, 1000]; the average area of ​​blind spots is a continuous value; and the alarm level is a discrete value, ranging from [1, 5]. Converting continuous data to discrete data facilitates association rule mining and improves its effectiveness.

[0077] S130. Perform association rule mining on the transactions to be mined to obtain the association rules between transaction items.

[0078] Specifically, association rules are patterns of correlation and dependency between variables in data. They are used to describe the relationships between different items in a given dataset, such as the relationship between the number of blind spots and the area of ​​blind spots. Association rule mining aims to discover frequent itemsets and association rules in data to determine the relationships and dependencies between these items.

[0079] S140, Obtain the initial decision tree for mining features.

[0080] Specifically, the initial decision tree includes a hierarchical order of feature mining. The decision tree is displayed in a branching tree structure, representing a mapping relationship between object attributes and object values. Each node in the tree represents an object, each branch path represents a possible attribute value, and each leaf node corresponds to the value of the object represented by the path from the root node to that leaf node.

[0081] like Figure 3As shown, for example, in the problem of insufficient radar detection area coverage, the root node is the mathematical representation of the radar blind zone problem (including the number of blind zones, the location of blind zones, and the size of blind zones, etc.), the first node is the number of radar blind zones, the second node is the area of ​​radar blind zones, and the third node is the size of radar blind zones, random selection of areas, selection of the nearest radar, and maximizing the coverage coefficient.

[0082] S150. Integrate the association rules into the initial decision tree to obtain the optimized decision tree.

[0083] Specifically, when integrating the discovered association rules into the decision tree, it is first necessary to find the node corresponding to the discovered rule (the position of the same association rule in the decision tree), and secondly, the value range of the discovered rule. These two are used as conditions to integrate the decision tree, so as to realize the automatic growth of the decision tree.

[0084] In one embodiment of this example, step S120 involves discretizing the data in the original data sample based on mining features to obtain transaction items corresponding to different mining features, and constructing a transaction to be mined based on different transaction items, including:

[0085] S200. Initialize the K cluster centers of the original data sample.

[0086] Specifically, the K-Means algorithm is used to perform cluster analysis on each column of data. First, K points in the feature space are randomly selected as the initial cluster centers.

[0087] S210. Calculate the spatial distance from each data point in the original data sample to the K cluster centers, and assign each data point to the cluster containing the nearest cluster center.

[0088] S220. Calculate the average distance from all data points in a cluster to the cluster center, and use the average distance as the new cluster center.

[0089] Specifically, for each other point, the distance to the K centers is calculated. For unknown points, the nearest cluster center is selected as the labeled category. Then, based on the labeled cluster center, the new center point (average value) of each cluster is recalculated. If the calculated new center point is the same as the original center point, then the process ends; otherwise, step S210 is repeated.

[0090] S230. Determine whether the new cluster center is the same as the original cluster center.

[0091] S240. If the new cluster center is the same as the original cluster center, then the clustering result for each mined feature is obtained.

[0092] Specifically, if the new cluster center is the same as the original cluster center, it means that the best clustering result has been obtained with the current K value.

[0093] S250. Based on the original data sample, the profile coefficient is calculated.

[0094] Specifically, to determine the optimal value of cluster K, the silhouette coefficient is used to evaluate the clustering effect. It is calculated by subtracting the difference between the distance of a sample to other samples in its own cluster and the sum of the distances to samples in other clusters outside its own cluster, and then dividing this difference by the larger of the two. The closer this value is to 1, the better the clustering effect; the closer this value is to -1, the worse the clustering effect; and the closer this value is to 0, the less effective the clustering. Let the number of clusters be k, and the clusters be p1, p2, ..., p... k Sample i belongs to cluster p i Then the silhouette coefficient of sample i is defined as:

[0095]

[0096] Among them, b i Represents the cohesion of sample i, and dist(i,j) is the distance between sample i and sample j, where i and j are samples within the same cluster. max{c i ,b i To find c i and b i The maximum value.

[0097]

[0098]

[0099] Among them, c i q represents the separation of sample i, where i and q are samples within different clusters.

[0100] The silhouette coefficient for overall clustering is:

[0101]

[0102] S260. Replace the K value based on the profile coefficient to obtain the optimal profile coefficient.

[0103] The optimal profile coefficient can be obtained by changing the K value based on the profile coefficient, including:

[0104] S2600, Calculate the profile coefficient for different K values.

[0105] S2610. Compare the contour coefficients under different K values, obtain the maximum contour coefficient, and take the maximum contour coefficient as the optimal contour coefficient.

[0106] Specifically, the silhouette coefficients under different K values ​​are calculated, and the largest silhouette coefficient is found. The change in K value is determined by the step size. The K value corresponding to the largest silhouette coefficient is the desired value. Clustering is then performed under the current K value to obtain the optimal clustering result. The optimal clustering result consists of different itemsets (transaction items), and multiple itemsets constitute the mining transaction.

[0107] S270. Obtain the optimal clustering result corresponding to the optimal silhouette coefficient, and use the optimal clustering result as the transaction item corresponding to different mining features.

[0108] In one embodiment of this example, step S130 involves performing association rule mining on the transactions to be mined, and obtaining the association rules between transaction items, including:

[0109] S300, Calculate the support level for each transaction item.

[0110] S310. Scan the transactions to be mined to obtain a candidate set.

[0111] S320. Based on candidate itemsets and support, frequent itemsets are obtained.

[0112] S330. Calculate the interest degree between different transaction items in the frequent item set to obtain the association rules between the transaction items.

[0113] Specifically, support is a metric used in data mining to measure the frequency or probability of an itemset appearing in the overall dataset. In association rule mining, support describes the prevalence or commonality of an itemset within the dataset. It can be used to determine whether an itemset has sufficient association relationships or as a basis for selecting frequent itemsets. Support can be calculated using the following formula: Support = Number of times the itemset appears / Size of the overall dataset.

[0114] The preprocessed dataset D can be represented as a set of n transactions, D = {t1,t2}. i+1 ,…,i n}, where each transaction t i ={tid i ,X i ,Y i}, where Y represents the meta-policy data item, |Y i | = 1, where X represents other environmental characteristic data items, |X i If |≥1, then the association rule of the decision tree is represented as the X→Y relationship. Apriori is a classic association rule mining algorithm used to discover frequent patterns.

[0115] Scanning the entire original data sample, the first candidate itemset is the set of all scanned items. Scanning involves iterating through every transaction in the original data sample; the set of all distinct transactions obtained is the candidate itemset. Next, we find the frequent (k+1)-itemset L. (k+1) First, concatenate the frequent k-itemsets with themselves to generate a candidate (k+1)-itemset, denoted as C. (k+1) The items are sorted in order, and items belonging to the same itemset are joined together. At the same time, frequent itemsets are generated based on support.

[0116] L k It is C k A subset of C, therefore, from the concept of a subset, we can know that C k Members may be frequent or infrequent, while L k All subsets must be in C k In the middle. Re-traverse the database, targeting C. k The frequency and weight of each item are obtained to calculate the support score. Itemsets with support scores that do not meet the requirements are deleted to determine L. k And the candidate set C k This could involve significant computational costs if the (k-1)-subset of a candidate k-itemset is not in the frequent itemset L. k-1 In the middle, then L k-1 The superset of a candidate k-itemset, i.e., the candidate k-itemset, is necessarily not frequent and can be directly defined in C. k Delete it. Repeat this process, traversing the database, until no new frequent itemsets are output, then terminate.

[0117] Using support and confidence as the basis for filtering association rules has certain drawbacks. If the support and confidence thresholds are too low, a large number of redundant or meaningless rules will be generated. Conversely, setting the thresholds too high may filter out some potential association rules that users are interested in. Therefore, this paper uses interest score to filter association rules. The specific calculation method is as follows:

[0118]

[0119] Where ial(X→Y) represents the interest level, supp(Y) represents the support level of transaction item Y, and conf(X→Y) represents the confidence level of the association rule "X→Y".

[0120] The interest measure satisfies ial(X→Y)∈[-1,1]. If the interest measure ial(X→Y)=0, it means that the itemsets X and Y in the pattern are independent of each other; if ial(X→Y)>0, it means that the itemsets X and Y in the pattern are positively correlated; if the interest measure ial(X→Y)<0, it means that the itemsets X and Y in the pattern are negatively correlated.

[0121] The association rule mining method using interest degree measurement requires that X→Y satisfy the following conditions for the rule to be valid:

[0122]

[0123] sp0 > 0, cf0 > 0, il0 > 0

[0124] Among them, sp0 and cf0 are hyperparameters that can be set manually, and supp(X∪Y) is the support level when X and Y appear simultaneously.

[0125] In one embodiment of this example, step S150, which involves incorporating association rules into the initial decision tree to obtain an optimized decision tree, includes:

[0126] S400, Obtain the mining feature level of the initial decision tree.

[0127] S410. Obtain the mining feature level and value range corresponding to the mining feature of the association rule.

[0128] S420. Based on the mining feature level and value range, the association rules are integrated into the initial decision tree to obtain the optimized decision tree.

[0129] S4200: Match the feature level to be mined with the feature level to be mined to obtain the merged node of the mined level.

[0130] S4201. Based on the integration nodes and value ranges, the association rules are integrated into the initial decision tree to obtain the optimized decision tree.

[0131] Specifically, the feature mining levels in the initial decision tree are determined by the order of the feature mining levels within the initial decision tree, such as... Figure 4 As shown.

[0132] Once an association rule X→Y is discovered, this rule should be reflected in the system's adaptive decision tree. Let's say one association rule X→Y is discovered, where X = {i1, i2, ..., i...}. t} represents t feature data items, Y = y s This represents a specific meta-policy s, where the order of data items in X conforms to the decision order of data attributes in the decision tree (feature mining hierarchy), and data item i... j This indicates that the attribute data takes values ​​within a specific range, i.e. (Value range).

[0133] The mining feature level of the discovered associated features is combined with the mining feature level of the initial decision tree and integrated into the same level position, that is, the integration node. The position to be integrated in the same level is found according to its value range.

[0134] Taking the problem of insufficient radar detection area coverage as an example, the initial decision tree mines features from top to bottom such as the number of blind spots, blind spot area, alarm level, and meta-policy number. After mining association rules, the optimized decision tree is obtained as follows: Figure 5 As shown in the figure, the red dashed lines represent the discovered association rules.

[0135] A decision tree growth system based on association rules, comprising:

[0136] The acquisition module is used to acquire raw data samples;

[0137] The extraction module is used to extract and mine features based on the original data samples;

[0138] The discrete module is used to discretize the data in the original data sample based on the mining features, obtain the transaction items corresponding to different mining features, and construct the transaction to be mined based on the different transaction items.

[0139] The mining module is used to mine association rules for the transactions to be mined, and to obtain the association rules between transaction items.

[0140] The second acquisition module is used to acquire the initial decision tree for mining features;

[0141] The building module is used to integrate association rules into the initial decision tree to obtain an optimized decision tree.

[0142] This application also discloses a terminal device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it uses a decision tree growth method based on association rules.

[0143] The terminal device can be a computer device such as a desktop computer, a laptop computer, or a cloud server. The terminal device includes, but is not limited to, a processor and a memory. For example, the terminal device may also include input / output devices, network access devices, and buses.

[0144] The processor can be a central processing unit (CPU). Of course, depending on the actual use, it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it in this regard.

[0145] The memory can be an internal storage unit of the terminal device, such as a hard disk or RAM of the terminal device, or an external storage device of the terminal device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), or flash memory card (FC) equipped on the terminal device. Furthermore, the memory can be a combination of internal storage units and external storage devices of the terminal device. The memory is used to store computer programs and other programs and data required by the terminal device. The memory can also be used to temporarily store data that has been output or will be output. This application does not limit this.

[0146] In this terminal device, the decision tree growth method based on association rules in the above embodiments is stored in the memory of the terminal device and loaded and executed on the processor of the terminal device for convenient use.

[0147] This application also discloses a computer-readable storage medium, which stores a computer program, wherein when the computer program is executed by a processor, it employs the decision tree growth method based on association rules described in the above embodiments.

[0148] The computer program can be stored in a computer-readable medium. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or certain middleware. The computer-readable medium includes any entity or device capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the computer-readable medium includes, but is not limited to, the above-mentioned components.

[0149] The association rule-based decision tree growth method described in the above embodiments is stored in the computer-readable storage medium and loaded and executed on the processor to facilitate the storage and application of the method.

[0150] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0151] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A decision tree growth method based on association rules, characterized in that, include: Obtain raw data samples; Based on the original data sample, extract and mine features; Based on the mining features, the data in the original data sample is discretized to obtain transaction items corresponding to different mining features, and different transaction items constitute the transaction to be mined; Perform association rule mining on the transactions to be mined to obtain the association rules between transaction items; Obtain the initial decision tree for the mined features; The association rules are then incorporated into the initial decision tree to obtain an optimized decision tree; The original data sample is data related to insufficient radar detection area coverage, including data sequence number, number of blind spots, alarm level data, and blind spot area data. The mining features are the features of the samples in the original sample, including the number of blind spots, blind spot area, and alarm level. The association rule mining performed on the transactions to be mined, to obtain the association rules between transaction items, includes: Calculate the support level for each transaction item; Scan the transactions to be mined to obtain a candidate set; Based on the candidate itemsets and the support, frequent itemsets are obtained; Calculate the degree of interest among different transaction items in the frequent itemset to obtain the association rules among the transaction items; The step of incorporating the association rules into the initial decision tree to obtain the optimized decision tree includes: Obtain the mining feature level of the initial decision tree; Obtain the mining feature level and value range corresponding to the mining features of the association rule; Based on the mining feature level and value range corresponding to the mining features of the association rule, the association rule is integrated into the initial decision tree to obtain an optimized decision tree; The process of integrating the association rules into the initial decision tree based on the mined feature level and value range to obtain the optimized decision tree includes: The mining feature level of the initial decision tree is matched with the mining feature level of the association rule to obtain the fusion node of the mining feature level; Based on the integration node and the value range, the association rule is integrated into the initial decision tree to obtain the optimized decision tree; The step of discretizing the data in the original data sample based on the mining features to obtain transaction items corresponding to different mining features includes: Initialize the K cluster centers of the original data sample; Calculate the spatial distance from each data point in the original data sample to the K cluster centers, and assign each data point to the cluster containing the nearest cluster center; Calculate the average distance from all data points in a cluster to the cluster center, and use the average distance as the new cluster center; Determine whether the new cluster center is the same as the original cluster center; If the new cluster center is the same as the original cluster center, then the clustering result for each of the mined features is obtained; Based on the original data sample, the contour coefficient is calculated; The optimal contour coefficient is obtained by changing the K value based on the contour coefficient. Obtain the optimal clustering result corresponding to the optimal contour coefficient, and use the optimal clustering result as the transaction item corresponding to different mining features.

2. The decision tree growth method based on association rules as described in claim 1, characterized in that, The step of changing the K value based on the contour coefficient to obtain the optimal contour coefficient includes: Calculate the profile coefficient for different K values; The contour coefficients under different K values ​​are compared to obtain the maximum contour coefficient, which is then taken as the optimal contour coefficient.

3. The decision tree growth method based on association rules as described in claim 1, characterized in that, The process of discretizing the data in the original data sample based on the mined features includes: Obtain the data types corresponding to the different mined features; Determine whether the data type is discrete data; If the data type is discrete data, then the data in the original data sample is discretized based on the mining features.

4. A decision tree growth system based on association rules, characterized in that, include: The acquisition module is used to acquire raw data samples; The extraction module is used to extract and mine features based on the original data sample; The discrete module is used to discretize the data in the original data sample based on the mining features to obtain transaction items corresponding to different mining features, and the different transaction items constitute the transaction to be mined; The mining module is used to perform association rule mining on the transaction to be mined, and obtain the association rules between transaction items; The second acquisition module is used to acquire the initial decision tree of the mined features; A construction module is used to integrate the association rules into the initial decision tree to obtain an optimized decision tree; The association rule mining performed on the transactions to be mined, to obtain the association rules between transaction items, includes: Calculate the support level for each transaction item; Scan the transactions to be mined to obtain a candidate set; Based on the candidate itemsets and the support, frequent itemsets are obtained; Calculate the degree of interest among different transaction items in the frequent itemset to obtain the association rules among the transaction items; The step of incorporating the association rules into the initial decision tree to obtain the optimized decision tree includes: Obtain the mining feature level of the initial decision tree; Obtain the mining feature level and value range corresponding to the mining features of the association rule; Based on the mining feature level and value range corresponding to the mining features of the association rule, the association rule is integrated into the initial decision tree to obtain an optimized decision tree; The process of integrating the association rules into the initial decision tree based on the mined feature level and value range to obtain the optimized decision tree includes: The mining feature level of the initial decision tree is matched with the mining feature level of the association rule to obtain the fusion node of the mining feature level; Based on the integration node and the value range, the association rule is integrated into the initial decision tree to obtain the optimized decision tree; The step of discretizing the data in the original data sample based on the mining features to obtain transaction items corresponding to different mining features includes: Initialize the K cluster centers of the original data sample; Calculate the spatial distance from each data point in the original data sample to the K cluster centers, and assign each data point to the cluster containing the nearest cluster center; Calculate the average distance from all data points in a cluster to the cluster center, and use the average distance as the new cluster center; Determine whether the new cluster center is the same as the original cluster center; If the new cluster center is the same as the original cluster center, then the clustering result for each of the mined features is obtained; Based on the original data sample, the contour coefficient is calculated; The optimal contour coefficient is obtained by changing the K value based on the contour coefficient. Obtain the optimal clustering result corresponding to the optimal contour coefficient, and use the optimal clustering result as the transaction item corresponding to different mining features.

5. A terminal device, comprising a memory and a processor, characterized in that, The memory stores a computer program that can run on a processor, and when the processor loads and executes the computer program, it employs the growth method according to any one of claims 1 to 3.

Images