CN110825615A - Software defect prediction method and system based on network embedding - Google Patents
Software defect prediction method and system based on network embedding Download PDFInfo
- Publication number
- CN110825615A CN110825615A CN201910898643.5A CN201910898643A CN110825615A CN 110825615 A CN110825615 A CN 110825615A CN 201910898643 A CN201910898643 A CN 201910898643A CN 110825615 A CN110825615 A CN 110825615A
- Authority
- CN
- China
- Prior art keywords
- software
- classes
- network
- defect prediction
- dependency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3604—Software analysis for verifying properties of programs
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Stored Programmes (AREA)
Abstract
The invention provides a software defect prediction method and a system based on network embedding. The method automatically excavates the potential structural characteristics in the software by utilizing a network embedding method, thereby more effectively predicting the software defects, solving the problem that the structural information in the software is not utilized in the original defect prediction method, improving the accuracy of a defect prediction model, helping developers to discover the defects possibly existing in the software in advance, reasonably distributing test tasks, reducing the test quantity and improving the efficiency of the testers.
Description
Technical Field
The invention relates to the field of software testing and defect prediction in software engineering, in particular to a software defect prediction method and system based on network embedding.
Background
The software defect prediction is to use the model established by the prior software data as the basis to carry out the defect prediction on the new code. By predicting the defects of the software, developers can be helped to find the defects possibly existing in the software in advance, the testing amount is reduced, and the testing efficiency is improved.
The current research aiming at software defect prediction technology can be divided into two stages, wherein the first stage is a feature extraction stage, and the defect representation is more efficient by manually designing new features or combining features; the second stage is a machine learning method classification stage, which provides a more effective prediction by using a new machine learning algorithm to build a more accurate model.
In the first stage, researchers have manually designed a number of features to distinguish defective and non-defective files from programs. For example, Halstead features based on the number of operands and operators; the complexity of the McCabe loop based on a program topological structure; the CK measurement element has the functions of inheritance, coupling, cohesion and the like in the object-oriented program and measures each attribute respectively; the MOOD measurement gives six measurement indexes based on object-oriented software from four aspects of inheritance, encapsulation, polymorphism and coupling; there are also code-based alteration features and some other object-oriented features.
In the second stage, many classification algorithms in machine learning have also been used in defect prediction, and the prediction model can predict whether a software module is defective only by using the feature values of the software module. The existing defect prediction classification technology comprises a support vector machine, naive Bayes, K neighbor, BP neural network and the like. The software reliability evaluation method can help testers to objectively evaluate the software reliability, and is beneficial to reasonable distribution of test resources so as to reduce the workload of the testers.
However, there is a clear dependency relationship between programs, the traditional metrics of these structural information cannot be captured, and the traditional metrics cannot distinguish code segments with different structures, so that the prediction effect is often not ideal.
Disclosure of Invention
In order to solve the above problems, the present invention provides a software defect prediction method and system based on Network Embedding, which utilizes a Network Embedding (Network Embedding) manner to mine the structural features of software, so as to achieve the purpose of performing software defect prediction analysis, and is used to assist software testing work. The directed graph nodes formed by the dependency relationship among the software are projected into the low-dimensional vector space in a network embedding mode, so that the structural features of the software are mined, and the problem that the structural features of the software cannot be captured by the traditional measurement features is effectively solved.
In order to achieve the purpose of the complaint, the invention adopts the specific technical scheme that:
a software defect prediction method based on network embedding comprises the following steps:
1) analyzing the source codes of classes in each file in the software project, and extracting a dependency relationship set between software classes;
2) constructing a dependency network among directed software classes according to the dependency relationship set among the software classes;
3) network embedding is carried out on the dependence network among the directed software classes to obtain software structure characteristic data;
4) and inputting the software structure feature vector data into a software defect prediction model and evaluating to obtain software defect prediction in a software project.
Further, the construction process of the software defect prediction model comprises the following steps:
1) analyzing the source codes of classes in each file in the training set, and extracting a dependency relationship set between software classes;
2) constructing a dependency network among directed software classes according to the dependency relationship set among the software classes;
3) network embedding is carried out on the dependence network among the directed software classes to obtain software structure characteristic data, and the software structure characteristic data is trained to obtain a software structure characteristic vector set;
4) and inputting the software structure characteristic vector set into a classifier for training to obtain a software defect prediction model based on network embedding.
Further, a Code Dependency collector method in a Dependency Finder is used for extracting the set of dependencies among the software classes.
Further, constructing the dependency network between the directed software classes through a HashMap (HashMap) structure in Java; the inter-directed software inter-class dependency network takes classes as nodes and the dependency relationship between the classes as connecting edges.
Further, the network embedding method is a random walk method in deep walk.
Further, carrying out classification balance processing on the software structure characteristic data, so that the final input training data is balanced; the method for carrying out the classification balance processing on the software structure characteristic data comprises the steps of reducing sample examples without defects in the software defect prediction data set and repeatedly training the sample examples with defects in the software defect prediction data set.
Further, the method for training the software structure characteristic data comprises the steps of sending the software structure characteristic data into a Skip-Gram model or a CBOW model; the learning training time of the Skip-Gram model is reduced by using a Hierarchica Softmax (hierarchy-based Softmax) model with a Huffman tree as an output mode.
Further, the classifier comprises a Logistic regression classifier and a random forest.
Further, the evaluation method is to calculate the harmonic mean of the software defect prediction accuracy and the recall rate.
A network embedding based software bug prediction system, comprising:
1) the software analysis module is used for analyzing the source codes of the classes in the files in the software project and extracting the dependency relationship among the software classes;
2) the dependent network construction module is used for constructing a dependent network between directed software classes according to the extracted dependent relationship between the software classes;
3) the characteristic data extraction module is used for carrying out network embedding on the dependence network among the directed software classes to obtain the software structure characteristic data;
4) and the software defect prediction module is used for inputting the software structure feature vector data into a software defect prediction model and evaluating the software structure feature vector data to obtain software defect prediction in a software project.
Compared with the existing software defect prediction method, the method has the following advantages:
the method automatically excavates the potential structural characteristics in the software by utilizing a network embedding method, thereby more effectively predicting the software defects, solving the problem that the structural information in the software is not utilized in the original defect prediction method and further improving the accuracy of a defect prediction model.
The invention aims to predict the defects of the software, help developers to find the defects possibly existing in the software in advance, reasonably distribute test tasks, reduce test quantity and improve the efficiency of testers.
Drawings
FIG. 1 is a flow chart of the construction of a software bug prediction model based on network embedding according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating extraction of dependencies between software classes according to an embodiment of the present invention.
FIG. 3 is a schematic diagram illustrating the generation of a dependency network between directed software classes in an embodiment of the present invention.
Fig. 4 is a flow chart of network embedding according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating a data set balancing process in accordance with an embodiment of the present invention.
FIG. 6 is a diagram illustrating regression model training in an embodiment of the present invention.
Detailed Description
In order to make the technical solutions in the embodiments of the present invention better understood and make the objects, features, and advantages of the present invention more comprehensible, the technical core of the present invention is described in further detail below with reference to the accompanying drawings and embodiments.
In the present invention, a software defect prediction method based on network embedding is proposed, which can effectively solve the above existing problems, and the method comprises the following steps:
as shown in fig. 1, the flow chart of the software defect prediction model based on network embedding is divided into five steps: firstly analyzing the dependency relationship among source code classes, then constructing a directional dependency network among software classes, then embedding the network by using Deepwalk on the constructed dependency network to obtain an embedded software structure characteristic vector set, balancing the software structure characteristic, and finally inputting the software structure characteristic into a Logistic regression classifier for training to obtain a software defect prediction model based on network embedding.
As shown in fig. 2, the Dependency relationship in the software is extracted by the Code Dependency collector method in the Dependency Finder, and three explicit Dependency relationships can be detected by the software source Code: function-to-function dependencies, function-to-class dependencies, class-to-class dependencies. The method only extracts the dependency between classes, namely extracts the dependency between each class in the software project and other classes in the project to obtain the set of the dependency between the software classes.
As shown in fig. 3, a dependency network between directed software classes is constructed by a HashMap structure in Java, and after the construction is completed, an adjacency matrix of the network is output. For the software project Prj, setting the dependent network thereof as PrjgraphAnd (V, E), wherein each node V ∈ V represents a class of Prj, and E represents a dependency relationship between the classes. Let ciIs v isiClass (d) represents, PrjgraphIn is vi→vjE is E, then ciAnd cjThere is at least one dependency. And constructing a directed graph formed by the dependency relationship between the software classes, wherein the classes are used as nodes, and the dependency relationship between the classes is a continuous edge.
As shown in fig. 4, a dependency network between directed software classes is constructed for network embedding, and the specific process is as follows: simulating a random walk sequence with fixed length by random walk in a network embedding method Deepwalk,and migrating the surrounding neighbor nodes. We assume an origin u and a wandering length l, ciIndicating the ith node of the walk, let c0U. Then node ciThe distribution of equation (1) will be followed:
wherein pivxI.e., the transition probability between nodes v and x, Z is a normalization constant.
And (2) carrying out random walk on each node according to a formula (1), and taking a sequence generated after the random walk is finished as an input and sending the input into a Skip-Gram model or a CBOW model for learning and training, so that a directed graph formed by the dependency relationship between software classes is projected into a low-dimensional vector space. Skip-Gram is a statistical language model, often used in the field of natural language processing. The sequence generated by the random walk can be analogized to a sentence in natural language processing, and the nodes in the sequence can be analogized to a word, so that the network can be embedded and learned by utilizing the model. The model can establish a simple neural network, and mainly aims to learn a hidden layer weight matrix in the network, take a certain number of nodes on the left side and the right side of an input node in a sequence to form a sliding window w, then select a certain number of nodes from the sliding window as output nodes so as to form (input and output) node pairs, input all the obtained node pairs into the neural network, generate a probability distribution through self-learning of the neural network, namely represent the probability of each node appearing with the input node at the same time, thereby completing the process of network embedding and obtaining a software structure characteristic vector set.
Preferably, the time for learning the Skip-Gram model can be greatly reduced by adopting a mode of constructing a Huffman tree as an output by adopting a Hierarchical Softmax model.
As shown in fig. 5, in general, the defective samples belong to a few classes, and the non-defective samples are the majority, which is expressed as an imbalance of sample classification, and will have a certain effect on the training of the model, and thus the effect of the final model. There are two ways to balance the data set, one is by reducing the majority of classes in the training data, i.e., software defects predict sample instances in the data set that are defect free. The other is by repeating a few classes in the training data, i.e., instances of samples in the software defect prediction dataset that are defective. The training data is balanced in a second mode, and the final balanced software structure feature vector set is obtained by randomly and repeatedly sampling the defective sample class until the number of the defective sample instances and the number of the non-defective sample instances reach a relative balance state of about (1: 1).
As shown in fig. 6, the balanced software structure feature vector set is input into a Logistic regression model or a classifier such as a Random Forest (Random Forest) for training, and the Logistic regression model is a nonlinear regression method obtained by introducing a Logistic regression function based on a linear regression model. In the fields of data mining and the like, Logistic regression is also the preferred method for the binary classification task, and new samples are classified by estimating coefficients and deviation values according to samples of known classes. After the sample estimation coefficients in the model are determined, the trained model is the software defect prediction model based on network embedding.
The data set of this embodiment is derived from a project (trusted data set for the field of software engineering), is an Apache open source project developed based on Java language, provides the version of the project, the number of classes, and the like, and labels the defective classes. For the same project, the project can be divided into a training set and a test set according to different versions, the project of the old version is used as the training set, and the project of the new version is used as the test set. In order to ensure the accuracy of the evaluation effect, when unbalanced data is processed, only the training set is balanced, and the test set is not processed. The specific process of model training is as follows:
where m is the number of training samples.
The model training is mainly to learn through a training data set to obtain a sample estimation coefficient theta shown in a formula (2), specifically, a cost function of the formula (3) is calculated through the training data, and when the cost function obtains a minimum value, the sample estimation coefficient theta is the optimal value, so that a software defect prediction model based on network embedding is obtained.
As an alternative embodiment, the resulting software defect prediction model may be evaluated. In the evaluation of the model, the F-measure value is used, which is the harmonic mean of accuracy and recall, as shown in equation (4).
The accuracy and the recall rate are composed of true positive (true positive), false positive (false positive) and false negative (false negative), wherein true positive means the predicted number of instances where the defective file is actually defective, false positive means the predicted number of instances where the defective file is actually not defective, and false negative means the predicted number of instances where the non-defective file is actually defective. Accuracy and recall are a pair of contradictory measures. In general, the higher the accuracy, the lower the recall rate is; when the recall rate is high, the accuracy rate is often low. The F-measure value balances the two, if the obtained values of the accuracy rate and the recall rate are higher, and the other value is lower, the final F-measure value is also lower, and the F-measure value is higher only when the values of the two are higher, so that the extreme condition is avoided, and the model can be better and accurately evaluated.
Evaluation of the model was performed using the poi project as an example, where the poi-2.5 version was used as the training set, the poi-3.0 version was used as the test set, and the average number of files for both versions was 413. The extracted software structure characteristic data in the poi-2.5 is used for training, and the F-measure value of the trained model tested at the poi-3.0 is 0.7892, which is higher than that of other software Defect Prediction methods based on CNN (CNN methods: Li J, He P, Zhu J, et al. software Defect Prediction visual connected Neural Network [ C ]//2017 IEEE International Conference on software Quality, Reliability and Security (QRS). IEEE,2017.) so as to prove the effectiveness of the method based on Network embedding in the field of software Defect Prediction.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the embodiments have been described in detail for the present invention, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered in the claims of the present invention.
Claims (10)
1. A software defect prediction method based on network embedding comprises the following steps:
1) analyzing the source codes of classes in each file in the software project, and extracting a dependency relationship set between software classes;
2) constructing a dependency network among directed software classes according to the dependency relationship set among the software classes;
3) network embedding is carried out on the dependence network among the directed software classes to obtain software structure characteristic data;
4) and inputting the software structure feature vector data into a software defect prediction model and evaluating to obtain software defect prediction in a software project.
2. The method of claim 1, wherein the software defect prediction model is constructed by a process comprising the steps of:
1) analyzing the source codes of classes in each file in the training set, and extracting a dependency relationship set between software classes;
2) constructing a dependency network among directed software classes according to the dependency relationship set among the software classes;
3) network embedding is carried out on the dependence network among the directed software classes to obtain software structure characteristic data, and the software structure characteristic data is trained to obtain a software structure characteristic vector set;
4) and inputting the software structure characteristic vector set into a classifier for training to obtain a software defect prediction model based on network embedding.
3. The method of claim 1 or 2, wherein the set of inter-software class dependencies are extracted using a code dependency gathering method in a dependency searcher.
4. The method of claim 1 or 2, wherein constructing the inter-directed software class dependency network is performed by a HashMap structure in Java; the inter-directed software inter-class dependency network takes classes as nodes and the dependency relationship between the classes as connecting edges.
5. The method of claim 1 or 2, wherein the method of network embedding is a random walk method in deep walk.
6. The method of claim 2, wherein the software structural feature data is class balanced such that the final input trained data is balanced; the method for carrying out the classification balance processing on the software structure characteristic data comprises the steps of reducing sample examples without defects in the software defect prediction data set and repeatedly training the sample examples with defects in the software defect prediction data set.
7. A method according to claim 2, wherein the software structural feature data is trained by feeding the software structural feature data into a Skip-Gram model or a CBOW model; and reducing the learning training time of the Skip-Gram model by using a Hiffman tree as an output mode.
8. The method of claim 2, wherein the classifier comprises a Logistic regression classifier, a random forest.
9. The method of claim 1, wherein the evaluation method is calculating a harmonic mean of software bug prediction accuracy and recall.
10. A network embedding based software bug prediction system, comprising:
1) the software analysis module is used for analyzing the source codes of the classes in the files in the software project and extracting the dependency relationship among the software classes;
2) the dependent network construction module is used for constructing a dependent network between directed software classes according to the extracted dependent relationship between the software classes;
3) the characteristic data extraction module is used for carrying out network embedding on the dependence network among the directed software classes to obtain the software structure characteristic data;
4) and the software defect prediction module is used for inputting the software structure feature vector data into a software defect prediction model and evaluating the software structure feature vector data to obtain software defect prediction in a software project.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910898643.5A CN110825615A (en) | 2019-09-23 | 2019-09-23 | Software defect prediction method and system based on network embedding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910898643.5A CN110825615A (en) | 2019-09-23 | 2019-09-23 | Software defect prediction method and system based on network embedding |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110825615A true CN110825615A (en) | 2020-02-21 |
Family
ID=69548128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910898643.5A Pending CN110825615A (en) | 2019-09-23 | 2019-09-23 | Software defect prediction method and system based on network embedding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110825615A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506504A (en) * | 2020-04-13 | 2020-08-07 | 扬州大学 | Software development process measurement-based software security defect prediction method and device |
| CN112115045A (en) * | 2020-08-19 | 2020-12-22 | 北京航空航天大学 | Failure prediction method for complex software system |
| CN112286807A (en) * | 2020-10-28 | 2021-01-29 | 北京航空航天大学 | Software defect positioning system based on source code file dependency relationship |
| CN113778862A (en) * | 2021-08-23 | 2021-12-10 | 中国人民解放军战略支援部队信息工程大学 | Software defect prediction method based on long-short term memory network and LASSO algorithm |
| CN114816997A (en) * | 2022-03-29 | 2022-07-29 | 湖北大学 | Defect prediction method based on graph neural network and bidirectional GRU feature extraction |
| CN116627845A (en) * | 2023-07-20 | 2023-08-22 | 山东鲁商通科技有限公司 | Capture-Recapture-based software defect prediction method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598375A (en) * | 2014-11-28 | 2015-05-06 | 江苏苏测软件检测技术有限公司 | Failure prediction method for software development |
| JP2016024477A (en) * | 2014-07-16 | 2016-02-08 | 株式会社Ｓｃｒｅｅｎホールディングス | Software defect prediction device, software defect prediction method, and software defect prediction program |
| CN105808435A (en) * | 2016-03-08 | 2016-07-27 | 北京理工大学 | Construction method of software defect evaluation model on the basis of complex network |
| CN107665172A (en) * | 2017-10-20 | 2018-02-06 | 北京理工大学 | A kind of Software Defects Predict Methods based on complicated weighting software network |
| CN108345544A (en) * | 2018-03-27 | 2018-07-31 | 北京航空航天大学 | A kind of software defect distribution analysis of Influential Factors method based on complex network |
| CN108829607A (en) * | 2018-07-09 | 2018-11-16 | 华南理工大学 | A kind of Software Defects Predict Methods based on convolutional neural networks |
-
2019
- 2019-09-23 CN CN201910898643.5A patent/CN110825615A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016024477A (en) * | 2014-07-16 | 2016-02-08 | 株式会社Ｓｃｒｅｅｎホールディングス | Software defect prediction device, software defect prediction method, and software defect prediction program |
| CN104598375A (en) * | 2014-11-28 | 2015-05-06 | 江苏苏测软件检测技术有限公司 | Failure prediction method for software development |
| CN105808435A (en) * | 2016-03-08 | 2016-07-27 | 北京理工大学 | Construction method of software defect evaluation model on the basis of complex network |
| CN107665172A (en) * | 2017-10-20 | 2018-02-06 | 北京理工大学 | A kind of Software Defects Predict Methods based on complicated weighting software network |
| CN108345544A (en) * | 2018-03-27 | 2018-07-31 | 北京航空航天大学 | A kind of software defect distribution analysis of Influential Factors method based on complex network |
| CN108829607A (en) * | 2018-07-09 | 2018-11-16 | 华南理工大学 | A kind of Software Defects Predict Methods based on convolutional neural networks |
Non-Patent Citations (1)
| Title |
|---|
| YU QU等: "node2defect: Using Network Embedding to Improve Software Defect Prediction", 《IEEE/ACM INTERNATIONAL CONFERENCE ON AUTOMATED SOFTWARE ENGINEERING》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506504A (en) * | 2020-04-13 | 2020-08-07 | 扬州大学 | Software development process measurement-based software security defect prediction method and device |
| CN111506504B (en) * | 2020-04-13 | 2023-04-07 | 扬州大学 | Software development process measurement-based software security defect prediction method and device |
| CN112115045A (en) * | 2020-08-19 | 2020-12-22 | 北京航空航天大学 | Failure prediction method for complex software system |
| CN112286807A (en) * | 2020-10-28 | 2021-01-29 | 北京航空航天大学 | Software defect positioning system based on source code file dependency relationship |
| CN112286807B (en) * | 2020-10-28 | 2022-01-28 | 北京航空航天大学 | Software defect positioning system based on source code file dependency relationship |
| CN113778862A (en) * | 2021-08-23 | 2021-12-10 | 中国人民解放军战略支援部队信息工程大学 | Software defect prediction method based on long-short term memory network and LASSO algorithm |
| CN114816997A (en) * | 2022-03-29 | 2022-07-29 | 湖北大学 | Defect prediction method based on graph neural network and bidirectional GRU feature extraction |
| CN114816997B (en) * | 2022-03-29 | 2023-08-18 | 湖北大学 | Defect prediction method based on graph neural network and bidirectional GRU feature extraction |
| CN116627845A (en) * | 2023-07-20 | 2023-08-22 | 山东鲁商通科技有限公司 | Capture-Recapture-based software defect prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825615A (en) | Software defect prediction method and system based on network embedding | |
| CN107665172B (en) | Software defect prediction method based on complex weighted software network | |
| CN106250461A (en) | A kind of algorithm utilizing gradient lifting decision tree to carry out data mining based on Spark framework | |
| CN111352971A (en) | Bank system monitoring data anomaly detection method and system | |
| CN113779272A (en) | Data processing method, device and equipment based on knowledge graph and storage medium | |
| CN109542783B (en) | Extended finite-state machine test data generation method based on variable segmentation | |
| Levin et al. | The co-evolution of test maintenance and code maintenance through the lens of fine-grained semantic changes | |
| CN113239358A (en) | Open source software vulnerability mining method based on knowledge graph | |
| CN112783786B (en) | Method, apparatus, device, medium and program product for generating test cases | |
| CN112966714A (en) | Edge time sequence data anomaly detection and network programmable control method | |
| CN112364352A (en) | Interpretable software vulnerability detection and recommendation method and system | |
| CN111563103A (en) | Method and system for detecting data blood margin | |
| CN117056834A (en) | Big data analysis method based on decision tree | |
| Shanthini et al. | Analyzing the effect of bagged ensemble approach for software fault prediction in class level and package level metrics | |
| CN117725592A (en) | Intelligent contract vulnerability detection method based on directed graph annotation network | |
| CN112019403B (en) | Cross-platform automatic mining method and system for message protocol state machine of Internet of things | |
| CN111898134B (en) | Intelligent contract vulnerability detection method and device based on LSTM and BiLSTM | |
| Alghanim et al. | Software defect density prediction using deep learning | |
| CN111090679B (en) | Time sequence data representation learning method based on time sequence influence and graph embedding | |
| Ostrowski et al. | Knowledge-based software testing agent using evolutionary learning with cultural algorithms | |
| CN112346974A (en) | Cross-mobile application program instant defect prediction method based on depth feature embedding | |
| Rajbahadur et al. | Pitfalls analyzer: quality control for model-driven data science pipelines | |
| Gu et al. | Hierarchical attention network for interpretable and fine-grained vulnerability detection | |
| CN106816871B (en) | State similarity analysis method for power system | |
| CN112579436B (en) | Micro-service software architecture identification and measurement method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200221 |
|
| RJ01 | Rejection of invention patent application after publication |