CN114186644A - Defect report severity prediction method based on optimized random forest - Google Patents

Defect report severity prediction method based on optimized random forest Download PDF

Info

Publication number
CN114186644A
CN114186644A CN202111633840.8A CN202111633840A CN114186644A CN 114186644 A CN114186644 A CN 114186644A CN 202111633840 A CN202111633840 A CN 202111633840A CN 114186644 A CN114186644 A CN 114186644A
Authority
CN
China
Prior art keywords
severity
defect
random forest
defect report
type
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.)
Withdrawn
Application number
CN202111633840.8A
Other languages
Chinese (zh)
Inventor
陈翔
葛骅
陈雪娇
林浩
苏展
缪芸
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nantong University
Original Assignee
Nantong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nantong University filed Critical Nantong University
Priority to CN202111633840.8A priority Critical patent/CN114186644A/en
Publication of CN114186644A publication Critical patent/CN114186644A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06395—Quality analysis or management

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Educational Administration (AREA)
  • Evolutionary Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Development Economics (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Strategic Management (AREA)
  • Game Theory and Decision Science (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of software quality assurance, and particularly relates to a defect report severity degree prediction method based on optimized random forest. Firstly, collecting and downloading a historical defect report from a defect tracking system where a project is located, and preprocessing a downloaded data set to obtain a target defect report data set; and then optimizing the random forest model by using a Bayesian hyperparameter optimization method to obtain an optimal hyperparameter, and finally constructing a defect report severity prediction model according to the optimal hyperparameter. The method uses a random forest model, and has stronger model generalization capability due to the fact that a plurality of base classifiers are integrated in the random forest model; the optimal hyper-parameter can be found out in a preset value range by utilizing a Bayesian hyper-parameter optimization method, the generalization capability of the random forest model is further enhanced, and the prediction capability of the model is improved.

Description

Defect report severity prediction method based on optimized random forest
Technical Field
The invention belongs to the technical field of software quality assurance, and particularly relates to a defect report severity degree prediction method based on optimized random forest.
Background
In the software development process, software defects inevitably occur, the software defects influence the software quality and need to be repaired in time, and the repair of the software defects accounts for a large proportion in the software development life cycle. Therefore, improving the efficiency of software bug fixes is key to ensuring the quality of software. Currently, to address this problem, many large projects use software defect report tracking systems to record defect information in order to quickly locate and repair defects.
The severity of software bug reports mainly comprises seven levels, namely, blocker, critic, main, normal, minor, trivision and enhancement. In actual operation, firstly, due to the difference of expression modes among users, the software defects of the same type may be judged as severity degrees of different levels; secondly, when developers classify software defect reports manually, the subjectivity is high, and the efficiency is low, so people urgently need to realize the classification of the software defect reports by means of an automatic technology. Inspired by the success of machine learning in the field of prediction in recent years, researchers have applied machine learning techniques to the problem of identifying the severity of software bug reports.
The traditional prediction method has low accuracy in predicting the severity of the software defect report, and can cause developers to spend a great deal of time on software defects with low urgency degree, thereby causing great influence on the software quality. In contrast, the machine learning algorithm can effectively identify the severity of the software defect report, and greatly reduces the cost of software maintenance.
Disclosure of Invention
Aiming at the problems in the technology, the invention provides a method for predicting the severity of the defect report based on the optimized random forest, so that the severity of the software defect report can be predicted more accurately.
In order to achieve the purpose, the invention adopts the following technical scheme: a defect report severity prediction method based on optimized random forests comprises the following steps:
(1) and extracting two attributes of the project historical defect report through a defect tracking system of the project: description information Summary and Severity degree, assuming that n defect reports are collected, a set of defect reports R ═ R (R) is formed1，R2，...，Rn) Wherein the ith defect is reported as Ri＝<Summary，Severity>；
(2) Preprocessing the description information Summary and the Severity degree sensing in the set R of the defect reports to obtain a target defect report data set RL ═ RL (RL)1，RL2，...，RLn)，
Where the ith defect is reported as RLi＝<Presummary，Preseverity＞，PresummaryRepresenting preprocessed description information, PreseverityIndicating the severity of the pre-treatment;
(3) dividing the target defect report data set RL into a training set and a verification set according to the proportion of 7: 3;
(4) and according to preset hyper-parameters, the hyper-parameters comprise: constructing a random forest model by using a training set based on the number of classifiers and the depth of a classification tree;
(5) predicting by using a verification set and the random forest model, and calculating an average absolute error;
(6) obtaining a better hyper-parameter by using a Bayesian hyper-parameter optimization method according to the preset hyper-parameter and the average absolute error, and taking the better hyper-parameter as the next preset hyper-parameter;
(7) repeating the steps (4) - (6) for K times to obtain the optimal hyper-parameter;
(8) and constructing a defect report severity prediction model based on the optimal hyper-parameters and the training set.
Further, as a preferred technical solution of the present invention, the step (2) specifically includes the steps of:
(2-1) deleting data with the Severity as normal in the set R of the defect report to obtain a data set T; (2-2) severity of Defect reportIs 7 types, and comprises the following components from low to high: the system comprises a first type enhancement, a second type trivision, a third type minor, a fourth type normal, a fifth type major, a sixth type critic and a seventh type blocker; since the fourth type normal severity is not considered, a classification determination is made for the remaining 6 defect report severity; determining the category of the severity of the defect reports of the first category of enhancement, the second category of trivision, the third category of minor, the fifth category of major, the sixth category of critic and the seventh category of block; if the severity of the defect report is a seventh class of blocker, setting the type dereferencing degree to be 1; if the severity is the sixth critical, setting the type value to be 2; if the severity is the fifth type major, setting the type value to be 3; if the severity is a third minor, setting the type value to be 4; if the severity is the second type of trivision, setting the type value to be 5; if the severity is the first type enhancement, setting the type value to be 6; obtaining the severity Pre after pretreatmentseverity；
(2-3) cutting description information Summary in the data set T into words, converting capital letters in the words into lowercase letters, deleting symbols in the words, removing stop words in the words based on a stop word list, and performing morphological restoration on the rest words based on spaces;
(2-4) representing the description information Summary in the data set T as a distributed vector based on a Skip-gram word embedding model to obtain the preprocessed description information Presummary；
(2-5) according to the severity Pre after the pretreatmentseverityAnd said preprocessed description information PresummaryAnd obtaining a target defect report data set.
Further, as a preferred technical solution of the present invention, the step (4) specifically includes the following steps:
(4-1) presetting a group of hyper-parameters, wherein the hyper-parameters comprise: the number of base classifiers and the depth of the classification tree;
(4-2) performing sampling with putting back on the training set by adopting a bagging method, wherein data which are not sampled are called as data outside a bag;
(4-3) selecting characteristic attributes by using a Gini coefficient as a splitting rule of a classification tree in a random forest model and using an out-of-bag error rate, wherein the out-of-bag error rate is expressed as:
Figure BDA0003441893620000031
(4-4) training and constructing the classification tree according to the data obtained by the sampling with the replacement based on the splitting rule and the error rate outside the bag;
(4-5) repeating (4-2) to (4-4) l times, and establishing l different base classifiers;
and (4-6) combining the l different base classifiers by using a majority voting method to construct a random forest model.
Further, as a preferred technical solution of the present invention, the step (5) specifically includes the steps of:
(5-1) predicting by using a verification set and the random forest model to obtain predicted values of the severity of each defect report;
(5-2) calculating an average absolute error according to the actual value of the severity of each defect report, wherein the average absolute error calculation formula is as follows:
Figure BDA0003441893620000032
in equation (1), MAE is the mean absolute error, m is the number of defect reports contained in the verification set, yiReporting the actual value of severity for the ith defect,
Figure BDA0003441893620000033
reporting severity prediction for the ith defect with a variance of [0, Q-1 ]]And Q is the number of severity levels of defect reports.
Compared with the prior art, the defect report severity prediction method based on the optimized random forest has the following technical effects:
a plurality of base classifiers are integrated in the random forest model, so that the model generalization capability is strong; the optimal hyper-parameter can be found out in a preset value range by utilizing a Bayesian hyper-parameter optimization method, the generalization capability of the random forest model is further enhanced, and the prediction capability of the model is improved.
Drawings
FIG. 1 is a schematic flow diagram of the present invention.
Detailed Description
The present invention will be further explained with reference to the drawings so that those skilled in the art can more deeply understand the present invention and can carry out the present invention, but the present invention will be explained below by referring to examples, which are not intended to limit the present invention.
Referring to FIG. 1, which is a flow chart illustrating the present invention, step (1) of the present invention collects a set R of 9972 defect reports by selecting two attributes of the descriptive information Summary and the Severity Severity of the firefox product defect report in mozilla project through the defect tracking system, wherein the ith defect report is Ri＝<Summary，Severity>The set of defect reports, R samples, is shown in Table 1;
TABLE 1 collective examples of Defect reports
Figure BDA0003441893620000041
Step (2) preprocessing the description information summery and the Severity degree coverage in the set R of the defect reports to obtain a target defect report data set RL (RL)1，RL2，...，RLn) Wherein the ith defect is reported as RLi＝<Presummary，Preseverity＞，PresummaryRepresenting preprocessed description information, PreseverityIndicating the severity after pretreatment, wherein the pretreatment steps are as follows:
(2-1) deleting data with the Severity as normal in the set R of the defect report to obtain a data set T; because the value is the default value of the severity of the defect report, the general defect report submitter selects the value when the severity of the defect report is not accurately known. Therefore, considering the defect report of the value when training the model introduces noise into the data set and reduces the prediction performance of the model.
(2-2), the severity of the defect report is generally in 7 categories, which are respectively from low severity to high severity: enhancement, trivision, minor, normal, major, critical, and block. Since normal severity was not considered, a classification was made for the remaining 6 defect report severity. Specifically, if the severity of the defect report is blocker, setting the type dereferencing degree to 1, if the severity is critic, setting the type dereferencing degree to 2, if the severity is major, setting the type dereferencing degree to 3, if the severity is minor, setting the type dereferencing degree to 4, if the severity is trivision, setting the type dereferencing degree to 5, if the severity is enhancement, setting the type dereferencing degree to 6, and obtaining the severity Pre-processed Preseverity；
(2-3) cutting description information Summary in the data set T into words, converting capital letters in the words into lowercase letters, deleting symbols in the words, removing stop words in the words based on a stop word list, and performing morphological restoration on the rest words based on spaces; the collective samples of defect reports in Table 1 are shown in Table 2 after this step;
TABLE 2 preliminary preprocessed target Defect report data set sample
Figure BDA0003441893620000051
(2-4) representing the description information Summary in the data set T as a distributed vector based on a Skip-gram word embedding model to obtain the preprocessed description information Presummary(ii) a The preliminary pre-processed target defect report data set example in table 2 is shown in table 3 after this step:
TABLE 3 preprocessed target Defect report data set sample
Figure BDA0003441893620000052
(2-5) according to the severity Pre after the pretreatmentseverityAnd said preprocessed description information PresummaryObtaining a target defect report data set;
(3) dividing the target defect report data set RL into a training set and a verification set according to the proportion of 7: 3;
(4) and according to preset hyper-parameters, the hyper-parameters comprise: constructing a random forest model by using a training set based on the number of classifiers and the depth of a classification tree; the step (4) specifically comprises the following steps:
(4-1) presetting a group of hyper-parameters, wherein the hyper-parameters comprise: the number of base classifiers and the depth of the classification tree;
(4-2) performing putting-back sampling on the training set by adopting a bagging method, wherein data which are not sampled are called out-of-bag data, and according to statistics, about 36.7% of data can be used as out-of-bag data;
(4-3) selecting characteristic attributes by using a Gini coefficient as a splitting rule of a classification tree in a random forest model and using an out-of-bag error rate, wherein the out-of-bag error rate is expressed as:
Figure BDA0003441893620000053
(4-4) training and constructing the classification tree according to the data obtained by the sampling with the replacement based on the splitting rule and the error rate outside the bag;
(4-5) repeating (4-2) to (4-4) l times, and establishing l different base classifiers;
and (4-6) combining the l different base classifiers by using a majority voting method to construct a random forest model.
(5) Predicting by using a verification set and the random forest model, and calculating an average absolute error; the step (5) specifically comprises the following steps:
(5-1) predicting by using a verification set and the random forest model to obtain predicted values of the severity of each defect report;
(5-2) calculating an average absolute error according to the actual value of the severity of each defect report, wherein the average absolute error calculation formula is as follows:
Figure BDA0003441893620000061
in equation (1), MAE is the mean absolute error, m is the number of defect reports contained in the verification set, yiReporting the actual value of severity for the ith defect,
Figure BDA0003441893620000062
reporting severity prediction for the ith defect with a variance of [0, Q-1 ]]And Q is the number of severity levels of defect reports.
(6) Obtaining a better hyperparameter according to the preset hyperparameter and the average absolute error by using a Bayesian hyperparameter optimization method, and taking the better hyperparameter as the next preset hyperparameter, wherein the parameters required by the Bayesian hyperparameter optimization method comprise:
(6-1), objective function: average absolute error in step (5);
(6-2) the parameter optimization space comprises: the number of base classifiers (between 10 and 100, with 5 steps), the depth of the classification tree (between 4 and 20, with 2 steps);
(6-3), optimizing algorithm: a Bayesian optimization algorithm based on a Gaussian process and a regression tree;
(7) repeating the steps (4) - (6) for K times to obtain the optimal hyper-parameter; in this embodiment, the user-defined K is 150, the obtained optimal result is that the number of base classes is 50, and the depth of the classification tree is 4;
(8) constructing a defect report severity prediction model based on the optimal hyper-parameters and the training set; according to the result of step (7), using the training set, a defect report severity prediction model can be constructed, which consists of 50 classification trees with a depth of 4.
The final prediction results for the target defect report data set examples in tables 1-3 are shown in table 4;
TABLE 4 comparison of target Defect report data set sample prediction results
Figure BDA0003441893620000063
In order to evaluate the performance of the defect report severity prediction model based on the optimized random forest, the MAE is used as an evaluation index, and the smaller the MAE value is, the smaller the prediction error of the model is. To illustrate the effects of the present invention, comparison was performed using a linear regression model as a reference. Both the invention and the linear regression model employ ten-fold cross validation on the training set.
The result shows that the MAF value of the prediction model is 0.52, the MAF value of the linear regression model is 0.89, and the prediction model is obviously superior to the linear regression model.
Firstly, collecting and downloading a historical defect report from a defect tracking system where a project is located, and preprocessing a downloaded data set to obtain a target defect report data set; and then optimizing the random forest model by using a Bayesian hyperparameter optimization method to obtain an optimal hyperparameter, and finally constructing a defect report severity prediction model according to the optimal hyperparameter. The method uses a random forest model, and has stronger model generalization capability due to the fact that a plurality of base classifiers are integrated in the random forest model; the optimal hyper-parameter can be found out in a preset value range by utilizing a Bayesian hyper-parameter optimization method, the generalization capability of the random forest model is further enhanced, and the prediction capability of the model is improved.
The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only illustrative of the present invention, and are not intended to limit the scope of the present invention, and any person skilled in the art should understand that equivalent changes and modifications made without departing from the concept and principle of the present invention should fall within the protection scope of the present invention.

Claims (4)

1. A method for predicting the severity of a defect report based on an optimized random forest is characterized by comprising the following steps:
(1) and extracting two attributes of the project historical defect report through a defect tracking system of the project: description information Summary and Severity degree, assuming that n defect reports are collected, a set of defect reports R ═ R (R) is formed1，R2，...，Rn) Wherein the ith defect is reported as Ri＝＜Summary，Severity＞；
(2) Preprocessing the description information Summary and the Severity degree sensing in the set R of the defect reports to obtain a target defect report data set RL ═ RL (RL)1，RL2，...，RLn)，
Where the ith defect is reported as RLi＝＜Presummary，Preseverity＞，PresummaryRepresenting preprocessed description information, PreseverityIndicating the severity of the pre-treatment;
(3) dividing the target defect report data set RL into a training set and a verification set according to the proportion of 7: 3;
(4) and according to preset hyper-parameters, the hyper-parameters comprise: constructing a random forest model by using a training set based on the number of classifiers and the depth of a classification tree;
(5) predicting by using a verification set and the random forest model, and calculating an average absolute error;
(6) obtaining a better hyper-parameter by using a Bayesian hyper-parameter optimization method according to the preset hyper-parameter and the average absolute error, and taking the better hyper-parameter as the next preset hyper-parameter;
(7) repeating the steps (4) - (6) for K times to obtain the optimal hyper-parameter;
(8) and constructing a defect report severity prediction model based on the optimal hyper-parameters and the training set.
2. The method for predicting the severity of defect report based on optimized random forest as claimed in claim 1, wherein said step (2) comprises the following steps:
(2-1) deleting data with the Severity as normal in the set R of the defect report to obtain a data set T;
(2-2), the severity of defect report is 7 types, from low to high: the system comprises a first type enhancement, a second type trivision, a third type minor, a fourth type normal, a fifth type major, a sixth type critic and a seventh type blocker; determining the category of the severity of the defect reports of the first category of enhancement, the second category of trivision, the third category of minor, the fifth category of major, the sixth category of critic and the seventh category of block; if the severity of the defect report is a seventh class of blocker, setting the type dereferencing degree to be 1; if the severity is the sixth critical, setting the type value to be 2; if the severity is the fifth type major, setting the type value to be 3; if the severity is a third minor, setting the type value to be 4; if the severity is the second type of trivision, setting the type value to be 5; if the severity is the first type enhancement, setting the type value to be 6; obtaining the severity Pre after pretreatmentseverity；
(2-3) cutting description information Summary in the data set T into words, converting capital letters in the words into lowercase letters, deleting symbols in the words, removing stop words in the words based on a stop word list, and performing morphological restoration on the rest words based on spaces;
(2-4) representing the description information Summary in the data set T as a distributed vector based on a Skip-gram word embedding model to obtain the preprocessed description information Presummary；
(2-5) according to the severity Pre after the pretreatmentseverityAnd said preprocessed description information PresummaryAnd obtaining a target defect report data set.
3. The method for predicting the severity of defect report based on optimized random forest as claimed in claim 1, wherein said step (4) comprises the following steps:
(4-1) presetting a group of hyper-parameters, wherein the hyper-parameters comprise: the number of base classifiers and the depth of the classification tree;
(4-2) performing sampling with putting back on the training set by adopting a bagging method, wherein data which are not sampled are called as data outside a bag;
(4-3) selecting characteristic attributes by using a Gini coefficient as a splitting rule of a classification tree in a random forest model and using an out-of-bag error rate, wherein the out-of-bag error rate is expressed as:
Figure FDA0003441893610000021
(4-4) training and constructing the classification tree according to the data obtained by the sampling with the replacement based on the splitting rule and the error rate outside the bag;
(4-5) repeating (4-2) to (4-4) l times, and establishing l different base classifiers;
and (4-6) combining the l different base classifiers by using a majority voting method to construct a random forest model.
4. The method for predicting the severity of defect report based on optimized random forest as claimed in claim 1, wherein said step (5) comprises the following steps:
(5-1) predicting by using a verification set and the random forest model to obtain predicted values of the severity of each defect report;
(5-2) calculating an average absolute error according to the actual value of the severity of each defect report, wherein the average absolute error calculation formula is as follows:
Figure FDA0003441893610000022
in equation (1), MAE is the mean absolute error, m is the number of defect reports contained in the verification set, yiReporting the actual value of severity for the ith defect,
Figure FDA0003441893610000023
reporting severity prediction for the ith defect with a variance of [0, Q-1 ]]And Q is the number of severity levels of defect reports.
CN202111633840.8A 2021-12-29 2021-12-29 Defect report severity prediction method based on optimized random forest Withdrawn CN114186644A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111633840.8A CN114186644A (en) 2021-12-29 2021-12-29 Defect report severity prediction method based on optimized random forest

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111633840.8A CN114186644A (en) 2021-12-29 2021-12-29 Defect report severity prediction method based on optimized random forest

Publications (1)

Publication Number Publication Date
CN114186644A true CN114186644A (en) 2022-03-15

Family

ID=80545122

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111633840.8A Withdrawn CN114186644A (en) 2021-12-29 2021-12-29 Defect report severity prediction method based on optimized random forest

Country Status (1)

Country Link
CN (1) CN114186644A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115048986A (en) * 2022-05-19 2022-09-13 河海大学 Ground surface freezing and thawing state classification method based on multi-classifier dynamic pruning selection

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115048986A (en) * 2022-05-19 2022-09-13 河海大学 Ground surface freezing and thawing state classification method based on multi-classifier dynamic pruning selection

Similar Documents

Publication Publication Date Title
CN111882446B (en) Abnormal account detection method based on graph convolution network
CN111914090B (en) Method and device for enterprise industry classification identification and characteristic pollutant identification
CN110991474A (en) Machine learning modeling platform
CN113221960B (en) Construction method and collection method of high-quality vulnerability data collection model
CN112682273B (en) Wind turbine generator fault detection method based on cost-sensitive lightweight gradient elevator
CN111949480A (en) Log anomaly detection method based on component perception
CN113887126A (en) Welding spot quality analysis method and device, terminal equipment and medium
CN111402236A (en) Hot-rolled strip steel surface defect grading method based on image gray value
CN114444620A (en) Indicator diagram fault diagnosis method based on generating type antagonistic neural network
CN115357764A (en) Abnormal data detection method and device
CN110287114B (en) Method and device for testing performance of database script
CN114254146A (en) Image data classification method, device and system
CN111507824A (en) Wind control model mold-entering variable minimum entropy box separation method
CN114186644A (en) Defect report severity prediction method based on optimized random forest
CN113177643A (en) Automatic modeling system based on big data
CN117541095A (en) Agricultural land soil environment quality classification method
CN111737993A (en) Method for extracting health state of equipment from fault defect text of power distribution network equipment
CN113704073B (en) Method for detecting abnormal data of automobile maintenance record library
CN116910526A (en) Model training method, device, communication equipment and readable storage medium
CN112306731B (en) Two-stage defect-distinguishing report severity prediction method based on space word vector
CN115081950A (en) Enterprise growth assessment modeling method, system, computer and readable storage medium
CN112148605B (en) Software defect prediction method based on spectral clustering and semi-supervised learning
CN114722960A (en) Method and system for detecting incomplete track of event log in business process
CN114490235A (en) Algorithm model for intelligently identifying quantity relation and abnormity of log data
CN114862092A (en) Evaluation method and device based on neural network

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20220315