CN109117380B - Software quality evaluation method, device, equipment and readable storage medium - Google Patents
Software quality evaluation method, device, equipment and readable storage medium Download PDFInfo
- Publication number
- CN109117380B CN109117380B CN201811144870.0A CN201811144870A CN109117380B CN 109117380 B CN109117380 B CN 109117380B CN 201811144870 A CN201811144870 A CN 201811144870A CN 109117380 B CN109117380 B CN 109117380B
- Authority
- CN
- China
- Prior art keywords
- evaluation
- target
- data
- neural network
- index system
- 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.)
- Active
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
- G06F11/3616—Software analysis for verifying properties of programs using software metrics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Stored Programmes (AREA)
Abstract
The invention discloses a software quality evaluation method, which comprises the following steps: acquiring an evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system; acquiring index data of target software according to a target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data; and processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software. According to the method, firstly, the initial evaluation index system is subjected to dimensionality reduction treatment to obtain a simplified but inaccurate target evaluation index system, and then the target evaluation index system is used for evaluating target software, so that the complexity and difficulty of an evaluation calculation process are reduced, and the software quality evaluation efficiency and accuracy are improved. Accordingly, the software quality evaluation device, the equipment and the readable storage medium disclosed by the invention also have the technical effects.
Description
Technical Field
The invention relates to the technical field of software quality detection, in particular to a software quality evaluation method, a device, equipment and a readable storage medium.
Background
Software quality is the sum of all features and all characteristics of a software product that are relevant to meeting the specified and implicit requirements capabilities. The software quality evaluation is to measure the quality characteristics, sub-characteristics, and the like of the software to be evaluated by using an appropriate technique and evaluate the measurement result, thereby giving an evaluation result of whether the software product can meet a specific requirement. The software quality evaluation can quantitatively give the quality level of the software, and the software can be correspondingly managed and improved according to the evaluation result of the software quality.
At present, the existing software quality evaluation methods are numerous and mainly include: the evaluation method comprises the steps of an analytic hierarchy process, a fuzzy comprehensive evaluation method, calculation based on a neural network and the like, wherein the evaluation methods are based on an original evaluation index system to acquire index data of software to be evaluated. Because the index parameters in the initial evaluation index system are many and complicated, the dimensionality is high, and redundant and useless index parameters are not avoided in the high-latitude index system, when the index data are collected based on the original evaluation index system, some redundant and useless index data can be collected, and the index data can interfere with the characteristic identification process, so that the accuracy of the evaluation result is reduced; in addition, when the target software is evaluated based on the neural network, if the neural network is trained based on the original evaluation index system, the training and verification processes of the neural network may be complicated, so that the complexity and difficulty of the calculation process can be improved, and the evaluation efficiency is affected.
Therefore, how to improve the efficiency and accuracy of software quality evaluation is a problem to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a software quality evaluation method, a device, equipment and a readable storage medium, so as to improve the software quality evaluation efficiency and accuracy.
In order to achieve the above purpose, the embodiment of the present invention provides the following technical solutions:
a software quality evaluation method comprises the following steps:
acquiring an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
acquiring index data of the target software according to the target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data;
and processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software.
Wherein, the evaluation neural network model is a wavelet neural network model.
Wherein, the construction process of the evaluation neural network model comprises the following steps:
acquiring initial training sample data and initial verification sample data based on the initial evaluation index system:
performing dimensionality reduction processing on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
determining a topological structure and parameter configuration of the evaluation neural network model, and training the evaluation neural network model by using the target training sample data;
inputting the target verification sample data into the evaluation neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by utilizing the trained evaluation neural network model.
Performing dimensionality reduction on the initial evaluation index system by using a Relieff algorithm to obtain a target evaluation index system, wherein the method comprises the following steps:
dividing index parameters in the initial evaluation index system into a plurality of samples;
determining a first nearest neighbor sample set that is homogeneous with each sample, and a second nearest neighbor sample set that is heterogeneous;
respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
deleting the weighted values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
The processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software includes:
and processing the evaluation data according to a basic probability assignment rule to obtain an evaluation result of the target software.
A software quality evaluation apparatus comprising:
the acquisition module is used for acquiring an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
the execution module is used for acquiring index data of the target software according to the target evaluation index system, inputting the index data into a preset neural network model and outputting evaluation data corresponding to the index data;
and the evaluation module is used for processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software.
Wherein, still include: an evaluation neural network model building module, the evaluation neural network model building module comprising:
the acquisition unit is used for acquiring initial training sample data and initial verification sample data based on the initial evaluation index system;
the dimensionality reduction unit is used for carrying out dimensionality reduction treatment on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
the training unit is used for determining the topological structure and the parameter configuration of the evaluation neural network model and training the evaluation neural network model by using the target training sample data;
the verification unit is used for inputting the target verification sample data into the evaluation neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by utilizing the trained evaluation neural network model.
Wherein the acquisition module comprises:
the dividing unit is used for dividing the index parameters in the initial evaluation index system into a plurality of samples;
a determining unit, configured to determine a first nearest-neighbor sample set that is homogeneous with each sample, and a second nearest-neighbor sample set that is heterogeneous with each sample;
the updating unit is used for respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
and the deleting unit is used for deleting the weight values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
A software quality evaluation apparatus comprising:
a memory for storing a computer program;
a processor for implementing the steps of the software quality evaluation method of any one of the above when executing the computer program.
A readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the software quality assessment method of any one of the above.
According to the scheme, the software quality evaluation method provided by the embodiment of the invention comprises the following steps: acquiring an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system; acquiring index data of the target software according to the target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data; and processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software.
According to the method, before the software quality is evaluated, the initial evaluation index system is subjected to dimensionality reduction to obtain a simplified target evaluation index system without inaccuracy, index data of target software is obtained according to the target evaluation index system, the obtained index data are processed through a preset evaluation neural network model, and finally the evaluation data output from the evaluation neural network model are processed according to a preset evaluation rule to obtain an evaluation result of the target software, so that the complexity and difficulty of an evaluation calculation process are reduced, and the software quality evaluation efficiency and accuracy are improved.
Accordingly, the software quality evaluation device, the equipment and the readable storage medium provided by the embodiment of the invention also have the technical effects.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a software quality evaluation method disclosed in the embodiments of the present invention;
FIG. 2 is a flow chart of another software quality evaluation method disclosed in the embodiments of the present invention;
FIG. 3 is a schematic diagram of a software quality evaluation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a software quality evaluation device disclosed in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a software quality evaluation method, a device, equipment and a readable storage medium, which aim to improve the software quality evaluation efficiency and accuracy.
Referring to fig. 1, a software quality evaluation method provided by an embodiment of the present invention includes:
s101, obtaining an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
in this embodiment, in order to reduce the complexity and data amount of the initial data, before the index data is collected, a ReliefF algorithm is used to perform dimension reduction processing on the initial evaluation index system. The Relieff algorithm can evaluate the classification capability of the index parameters in the initial evaluation index system, select the index parameters with strong classification capability, and remove invalid and redundant index parameters to reduce the dimension of the evaluation index system and obtain a target evaluation index system, so that the design of a neural network can be further simplified.
Specifically, the core idea of the dimension reduction processing is as follows: and measuring the correlation between the features and the categories by using the distance, endowing different weights to the features according to the correlation, selecting the features with strong correlation with the categories according to the weights, and removing the irrelevant and redundant features. Namely: and evaluating the classification capability of each feature in the initial evaluation index system by utilizing a RelifF algorithm, selecting the feature with stronger classification capability, and removing irrelevant and redundant features. Wherein, each characteristic can be regarded as each index parameter.
S102, acquiring index data of target software according to a target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data;
it should be noted that the evaluation neural network model is a wavelet neural network model.
The wavelet neural network model is used for carrying out relevant calculation, so that the problem that evaluation data is greatly influenced by expert subjective factors can be solved. The relation between the input and the output of the neural network model is a highly nonlinear mapping relation, and the relational expression is generally difficult to express, so that objective weight is automatically obtained according to feedback information through a training learning process, and the objectivity and the comprehensiveness of evaluation data are improved. It should be noted that the wavelet neural network model integrates the advantages of wavelet analysis and neural network model in processing nonlinear problems, and can solve the problems of blindness of the structural design of the BP neural network model and easiness in falling into local optimization.
And S103, processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software.
Therefore, the embodiment provides a software quality evaluation method, which includes performing dimensionality reduction on an initial evaluation index system before evaluating software quality to obtain a simplified target evaluation index system without inaccuracy, further acquiring index data of target software according to the target evaluation index system, processing the acquired index data through a preset evaluation neural network model, and finally processing evaluation data output from the evaluation neural network model according to a preset evaluation rule to obtain an evaluation result of the target software, so that complexity and difficulty of an evaluation calculation process are reduced, and software quality evaluation efficiency and accuracy are improved.
The embodiment of the invention discloses another software quality evaluation method, and compared with the previous embodiment, the technical scheme is further explained and optimized by the embodiment.
Referring to fig. 2, another software quality evaluation method provided by the embodiment of the present invention includes:
s201, obtaining an initial evaluation index system and target software, and performing dimensionality reduction processing on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
s202, acquiring index data of target software according to a target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data;
and S203, processing the evaluation data according to the basic probability assignment rule to obtain the evaluation result of the target software.
In this embodiment, in order to further improve the accuracy of the evaluation result, a plurality of wavelet neural network models may be used to process the index data of the target software, so as to obtain a plurality of evaluation data. Each evaluation neural network model can be processed to obtain evaluation data, and the topological structure and parameter configuration of each wavelet neural network are different. When each wavelet neural network model outputs evaluation data, the evaluation data can be output according to a preset local decision rule. Namely: when each wavelet neural network model outputs evaluation data, the evaluation data needs to be output according to a preset local decision criterion, and the local decision criterion can be flexibly adjusted according to actual application conditions, so that the specification is not specifically limited herein.
And then, fusing the plurality of evaluation data by utilizing the DS theory to obtain a fusion result, and taking the fusion result as the evaluation data corresponding to the index data, thereby avoiding a bias result brought by processing based on one evaluation neural network. The process of fusing the plurality of evaluation data includes: determining an identification frame of the DS evidence theory by utilizing the evaluation result grade in the target evaluation index system; generating basic probability assignment of each proposition in the recognition framework based on the plurality of evaluation data; and combining the Dempster combination rule and the basic probability assigned values of all the propositions in the identification frame to perform pairwise fusion on the plurality of evaluation data to obtain a fusion result.
And processing the obtained fusion result according to a basic probability assignment rule to obtain an evaluation result of the target software.
Therefore, according to the method, before software quality is evaluated, an initial evaluation index system is subjected to dimensionality reduction to obtain a simplified target evaluation index system without inaccuracy, index data of target software is obtained according to the target evaluation index system, the obtained index data are processed through a plurality of preset evaluation neural network models, obtained processing results are fused, and finally the fused result fused by the processing results is processed according to a preset evaluation rule to obtain the evaluation result of the target software, so that the complexity and difficulty of an evaluation calculation process are reduced, and the software quality evaluation efficiency and accuracy are improved.
Based on any of the above embodiments, it should be noted that the building process of the evaluation neural network model includes:
acquiring initial training sample data and initial verification sample data based on the initial evaluation index system;
performing dimensionality reduction processing on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
determining a topological structure and parameter configuration of the evaluation neural network model, and training the evaluation neural network model by using the target training sample data;
inputting the target verification sample data into the evaluation neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by utilizing the trained evaluation neural network model.
Based on any of the above embodiments, it should be noted that the performing, by using the ReliefF algorithm, the dimension reduction processing on the initial evaluation index system to obtain a target evaluation index system includes:
dividing index parameters in the initial evaluation index system into a plurality of samples;
determining a first nearest neighbor sample set that is homogeneous with each sample, and a second nearest neighbor sample set that is heterogeneous;
respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
deleting the weighted values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
In the following, a software quality evaluation device provided by an embodiment of the present invention is introduced, and a software quality evaluation device described below and a software quality evaluation method described above may be referred to each other.
Referring to fig. 3, an apparatus for evaluating software quality according to an embodiment of the present invention includes:
an obtaining module 301, configured to obtain an initial evaluation index system and target software, and perform dimension reduction processing on the initial evaluation index system by using a ReliefF algorithm to obtain a target evaluation index system;
the execution module 302 is configured to obtain index data of the target software according to the target evaluation index system, input the index data into a preset neural network model, and output evaluation data corresponding to the index data;
and the evaluation module 303 is configured to process the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software.
Wherein, still include: an evaluation neural network model building module, the evaluation neural network model building module comprising:
the acquisition unit is used for acquiring initial training sample data and initial verification sample data based on the initial evaluation index system;
the dimensionality reduction unit is used for carrying out dimensionality reduction treatment on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
the training unit is used for determining the topological structure and the parameter configuration of the evaluation neural network model and training the evaluation neural network model by using the target training sample data;
the verification unit is used for inputting the target verification sample data into the evaluation neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by utilizing the trained evaluation neural network model.
Wherein the acquisition module comprises:
the dividing unit is used for dividing the index parameters in the initial evaluation index system into a plurality of samples;
a determining unit, configured to determine a first nearest-neighbor sample set that is homogeneous with each sample, and a second nearest-neighbor sample set that is heterogeneous with each sample;
the updating unit is used for respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
and the deleting unit is used for deleting the weight values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
Wherein the evaluation module is specifically configured to:
and processing the evaluation data according to a basic probability assignment rule to obtain an evaluation result of the target software.
It can be seen that this embodiment provides a software quality evaluation device, including: the device comprises an acquisition module, an execution module and an evaluation module. Firstly, an acquisition module acquires an initial evaluation index system and target software, and the Relieff algorithm is utilized to perform dimensionality reduction processing on the initial evaluation index system to obtain a target evaluation index system; then the execution module acquires index data of the target software according to the target evaluation index system, inputs the index data into a preset neural network model and outputs evaluation data corresponding to the index data; and finally, the evaluation module processes the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software. Therefore, all modules are in work and cooperation and take their own roles, so that the complexity and difficulty of the evaluation and calculation process are reduced, and the software quality evaluation efficiency and accuracy are improved.
In the following, a software quality evaluation device provided by an embodiment of the present invention is introduced, and a software quality evaluation device described below and a software quality evaluation method and device described above may be referred to each other.
Referring to fig. 4, a software quality evaluation apparatus provided in an embodiment of the present invention includes:
a memory 401 for storing a computer program;
a processor 402 for implementing the steps of the software quality evaluation method according to any of the embodiments described above when executing the computer program.
In the following, a readable storage medium provided by an embodiment of the present invention is introduced, and a readable storage medium described below and a software quality evaluation method, device and apparatus described above may be referred to each other.
A readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the software quality assessment method according to any of the embodiments described above.
The following evaluation schemes were designed according to the methods provided in the present specification:
1. selecting a software quality evaluation index system, namely an initial evaluation index system; the software quality evaluation index system comprises: and (4) evaluating the quality of the software.
Setting the software quality evaluation index system as U ═ T1，T2，…，TMWhere T isi(i ═ 1, 2, …, M) data representing the i-th index; the software quality evaluation grade system is theta ═ V1，V2，…，VN}；
2. And collecting sample data according to the selected software quality evaluation index system.
3. The sample data is divided into a training sample set and a testing sample set.
The method comprises the steps of randomly dividing sample data into a training sample set and a testing sample set, and randomly selecting 70% of the sample data as the training sample set and 30% of the sample data as the testing sample set without loss of generality.
4. And performing dimensionality reduction on the software quality evaluation index system to obtain a new evaluation index system.
Let training sample set D (L × M), consist of L samples, each sample has M feature representations T1，T2，…，TMThe feature weight vector obtained after the RelifF algorithm processing is ω (1 × M), where ω (T)l) Represents a feature TlThe weight of (2). The training sample set consists of index parameters in a software quality evaluation index system.
The processing procedure of the RelifF algorithm is as follows:
(1) setting the initial value of the feature weight vector omega as a zero vector, setting the cycle number as gamma (gamma is less than or equal to L), and setting the sample class number as c (c is more than or equal to 2) in the sample set D.
(2) And (c) circularly executing the steps a to c for gamma times:
A. randomly selecting a sample Y from the sample set Di；
b. At YiIn the same class as YiNearest neighbor k samples Zj(j＝1，2，…，k)；
c. At YiFind the same as Y in each of the different classesiNearest neighbor k samples Xj(j＝1，2，…，k)；
And updating each characteristic weight according to the following formula:
wherein: class (Y)i) Represents a sample YiClass number of (2); diff (T)l，Y1，Y2) Represents a sample Y1And Y2The Euclidean distance of (1) is used for measuring the dissimilarity degree of two samples, and the calculation method is as follows:
for discrete features:
for the continuous feature:
wherein value (T)l，Y1) Represents a sample Y1At a characteristic TlThe value of (c).
From the above formula, it can be seen that: for a dimension feature TlIf it is good for classification, homogeneous samples should be brought close and heterogeneous samples should be brought far apart. Namely: two samples from the same class are characterized by TlDistance diff (T) abovel，Yi，Zj) The smaller and two samples from different classes are at TlDistance diff (T) abovel，Yi，Xj) The larger the weight ω (T) it obtainsl) The larger.
(3) And outputting the feature weight vector omega.
(4) The features are arranged in descending order according to the weight, and the weight higher than a threshold value gamma is selectedωAnd (4) eliminating the features with small weight value, thereby constructing a new feature set.
The new feature set is a new software quality evaluation index system after dimension reduction processing by utilizing a Relieff algorithm, and the new feature set is
5. And determining the topological structure and parameter configuration of the wavelet neural network.
(1) Topological structure:
inputting a layer: the number of layers is 1, the number of nodes and a new software quality evaluation index system U+The number of the middle indexes is the same, namely the number of the nodes is F.
② hidden layer: the number of layers is 1, and the node number is determined by combining an empirical formula and learning training results, where i is the number of nodes of the input layer, o o is the number of nodes of the output layer, a a is [1, 10 ]]An integer within the interval, wherein,
output layer: the number of layers is 1, the number of nodes is the same as the number of software quality grades in a software quality evaluation grade system theta, namely the number of nodes is N.
(2) Parameter configuration:
hidden layer transfer function: a Morlet wavelet function;
output layer transfer function: a sigmoid function;
target error set value: 0.01;
momentum factor value: 0.9;
learning rate: setting the initial learning rate to be 0.01 by adopting a self-adaptive learning victory ratio optimization algorithm;
training end conditions: the maximum number of training steps is 5000 steps, and the time is 30 seconds.
6. And training the wavelet neural network by using the training sample set.
The original training sample is evaluated according to a new software quality index system U+And setting the indexes to generate a new training sample, and training the wavelet neural network.
7. And determining a local decision criterion used for determining a software quality evaluation result according to the output of the wavelet neural network.
Let the output of the wavelet neural network be G ═ G1，g2，…，gNH, if g existspIf the following formula is satisfied, the software quality evaluation result is called Vp。
8. And testing the wavelet neural network by using the test sample set.
The original test sample is evaluated according to a new software quality index system U+And setting indexes in the wavelet neural network to generate a new test sample to test the wavelet neural network.
Inputting a test sample into a trained wavelet neural network, determining a software quality evaluation result aiming at the output of the wavelet neural network under the guidance of a decision criterion, verifying whether the software quality evaluation result given by the wavelet neural network is consistent with the real software quality evaluation result of the test sample, if so, selecting the next test sample for processing, otherwise, determining that the test is not passed, modifying the parameter configuration, and then re-training and testing. And if the software quality evaluation results of the current wavelet neural network for all the test samples in the test sample set are consistent with the real software quality evaluation results, the test is considered to be passed, and the wavelet neural network can be used for subsequent software quality evaluation.
9. And inputting the quality evaluation index data of the software to be evaluated into the trained wavelet neural network.
10. And under the guidance of a basic probability assignment rule, determining a software quality evaluation result of the software to be evaluated according to the output of the wavelet neural network.
In conclusion, the classification capability of each index parameter in the initial evaluation index system is evaluated by utilizing the Relieff algorithm, the index with strong classification capability is selected, invalid and redundant indexes are removed, the dimension of the evaluation index system is reduced, a new evaluation index system is obtained, and the calculation complexity of the wavelet neural network is further simplified.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (7)
1. A software quality evaluation method is characterized by comprising the following steps:
acquiring an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
acquiring index data of the target software according to the target evaluation index system, inputting the index data into a preset evaluation neural network model, and outputting evaluation data corresponding to the index data;
processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software;
wherein the evaluation neural network model is a wavelet neural network model; the topological structure of the wavelet neural network model comprises: the number of layers is 1 input layer; hidden layers with 1 layer number; the number of layers is 1 output layer; the parameter configuration of the wavelet neural network model comprises the following steps: hidden layer transfer function: a Morlet wavelet function; output layer transfer function: a sigmoid function; processing index data of the target software by using a plurality of wavelet neural network models to obtain a plurality of evaluation data;
wherein, the construction process of the evaluation neural network model comprises the following steps:
acquiring initial training sample data and initial verification sample data based on the initial evaluation index system;
performing dimensionality reduction processing on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
determining a topological structure and parameter configuration of the evaluation neural network model, and training the evaluation neural network model by using the target training sample data;
inputting the target verification sample data into the evaluation neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by utilizing the trained evaluation neural network model.
2. The software quality evaluation method according to claim 1, wherein the performing dimension reduction processing on the initial evaluation index system by using a Relieff algorithm to obtain a target evaluation index system comprises:
dividing index parameters in the initial evaluation index system into a plurality of samples;
determining a first nearest neighbor sample set that is homogeneous with each sample, and a second nearest neighbor sample set that is heterogeneous;
respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
deleting the weighted values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
3. The software quality evaluation method according to claim 2, wherein the processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software comprises:
and processing the evaluation data according to a basic probability assignment rule to obtain an evaluation result of the target software.
4. A software quality evaluation apparatus, comprising:
the acquisition module is used for acquiring an initial evaluation index system and target software, and performing dimensionality reduction on the initial evaluation index system by utilizing a Relieff algorithm to obtain a target evaluation index system;
the execution module is used for acquiring index data of the target software according to the target evaluation index system, inputting the index data into a preset neural network model and outputting evaluation data corresponding to the index data;
the evaluation module is used for processing the evaluation data according to a preset evaluation rule to obtain an evaluation result of the target software;
wherein the neural network model is a wavelet neural network model; the topological structure of the wavelet neural network model comprises: the number of layers is 1 input layer; hidden layers with 1 layer number; the number of layers is 1 output layer; the parameter configuration of the wavelet neural network model comprises the following steps: hidden layer transfer function: a Morlet wavelet function; output layer transfer function: a sigmoid function; processing index data of the target software by using a plurality of wavelet neural network models to obtain a plurality of evaluation data;
wherein, still include: a neural network model building module, the neural network model building module comprising:
the acquisition unit is used for acquiring initial training sample data and initial verification sample data based on the initial evaluation index system;
the dimensionality reduction unit is used for carrying out dimensionality reduction treatment on the initially acquired training sample data and the initially verified sample data based on the target evaluation index system to obtain target training sample data and target verified sample data;
the training unit is used for determining the topological structure and the parameter configuration of the neural network model and training the neural network model by using the target training sample data;
the verification unit is used for inputting the target verification sample data into the neural network model after training is finished, and judging whether the output evaluation data is consistent with the original evaluation data of the target verification sample data; and if so, evaluating the index data of the target software by using the trained neural network model.
5. The software quality evaluation device according to claim 4, wherein the acquisition module includes:
the dividing unit is used for dividing the index parameters in the initial evaluation index system into a plurality of samples;
a determining unit, configured to determine a first nearest-neighbor sample set that is homogeneous with each sample, and a second nearest-neighbor sample set that is heterogeneous with each sample;
the updating unit is used for respectively updating the weight of the characteristic value in each first sample set and the weight of the characteristic value in each second sample set by using a preset formula to obtain a weight set;
and the deleting unit is used for deleting the weight values smaller than a preset threshold value in the weight set to obtain a target weight set, and determining the target evaluation index system according to the target weight set.
6. A software quality evaluation apparatus characterized by comprising:
a memory for storing a computer program;
a processor for implementing the steps of the software quality assessment method according to any one of claims 1 to 3 when executing said computer program.
7. A readable storage medium, characterized in that the readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps of the software quality assessment method according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811144870.0A CN109117380B (en) | 2018-09-28 | 2018-09-28 | Software quality evaluation method, device, equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811144870.0A CN109117380B (en) | 2018-09-28 | 2018-09-28 | Software quality evaluation method, device, equipment and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109117380A CN109117380A (en) | 2019-01-01 |
| CN109117380B true CN109117380B (en) | 2022-02-08 |
Family
ID=64857308
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811144870.0A Active CN109117380B (en) | 2018-09-28 | 2018-09-28 | Software quality evaluation method, device, equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109117380B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110059118B (en) * | 2019-04-26 | 2022-04-08 | 迪爱斯信息技术股份有限公司 | Weight calculation method and device of characteristic attribute and terminal equipment |
| CN110221976B (en) * | 2019-05-28 | 2023-08-22 | 广西电网有限责任公司电力科学研究院 | Quantitative evaluation method for quality of metering terminal software based on measurement technology |
| CN110968512B (en) * | 2019-11-29 | 2021-06-01 | 中国科学院长春光学精密机械与物理研究所 | Software quality evaluation method, device, equipment and computer readable storage medium |
| CN111026661B (en) * | 2019-12-06 | 2023-09-19 | 广东省科技基础条件平台中心 | Comprehensive testing method and system for software usability |
| CN111506510B (en) * | 2020-04-21 | 2023-08-08 | 腾讯科技（深圳）有限公司 | Software quality determining method and related device |
| CN111767212B (en) * | 2020-06-17 | 2024-05-14 | 中国船舶重工集团公司第七一九研究所 | Software code quality evaluation method, device, equipment and storage medium |
| CN112101566A (en) * | 2020-09-11 | 2020-12-18 | 石化盈科信息技术有限责任公司 | Prediction model training method, price prediction method, storage medium, and electronic device |
| CN113626070B (en) * | 2021-08-06 | 2023-10-31 | 上海浦东发展银行股份有限公司 | Method, device, equipment and storage medium for configuring code quality index |
| CN113806223A (en) * | 2021-09-10 | 2021-12-17 | 北京中联国成科技有限公司 | Software evaluation method and device |
| CN114418269A (en) * | 2021-11-30 | 2022-04-29 | 哈尔滨工业大学 | Industrial robot safety evaluation index construction method |
| CN114816963B (en) * | 2022-06-28 | 2022-09-20 | 南昌航空大学 | Embedded software quality evaluation method, system, computer and readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389598A (en) * | 2015-12-28 | 2016-03-09 | 中国石油大学(华东) | Feature selecting and classifying method for software defect data |
| CN107027023A (en) * | 2017-04-24 | 2017-08-08 | 北京理工大学 | VoIP based on neutral net is without reference video communication quality method for objectively evaluating |
| CN107766254A (en) * | 2017-11-13 | 2018-03-06 | 长春长光精密仪器集团有限公司 | A kind of Evaluation of Software Quality and system based on step analysis |
| CN107797931A (en) * | 2017-11-13 | 2018-03-13 | 长春长光精密仪器集团有限公司 | A kind of method for evaluating software quality and system based on second evaluation |
-
2018
- 2018-09-28 CN CN201811144870.0A patent/CN109117380B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389598A (en) * | 2015-12-28 | 2016-03-09 | 中国石油大学(华东) | Feature selecting and classifying method for software defect data |
| CN107027023A (en) * | 2017-04-24 | 2017-08-08 | 北京理工大学 | VoIP based on neutral net is without reference video communication quality method for objectively evaluating |
| CN107766254A (en) * | 2017-11-13 | 2018-03-06 | 长春长光精密仪器集团有限公司 | A kind of Evaluation of Software Quality and system based on step analysis |
| CN107797931A (en) * | 2017-11-13 | 2018-03-13 | 长春长光精密仪器集团有限公司 | A kind of method for evaluating software quality and system based on second evaluation |
Non-Patent Citations (2)
| Title |
|---|
| 基于人工神经网络的软件质量预测模型研究;胡求索;《中国优秀硕士学位论文全文数据库信息科技辑》;20071115;摘要、第四章 * |
| 特征选择新算法研究;冯宗翰;《中国优秀硕士学位论文全文数据库信息科技辑》;20110815;第4.1、5.1、5.2节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109117380A (en) | 2019-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109117380B (en) | Software quality evaluation method, device, equipment and readable storage medium | |
| CN113779272B (en) | Knowledge graph-based data processing method, device, equipment and storage medium | |
| JP7207790B2 (en) | Assembly error correction for assembly lines | |
| US20050049913A1 (en) | Method and apparatus for automated feature selection | |
| Kamada et al. | An adaptive learning method of restricted Boltzmann machine by neuron generation and annihilation algorithm | |
| CN110968512B (en) | Software quality evaluation method, device, equipment and computer readable storage medium | |
| CN114780831A (en) | Sequence recommendation method and system based on Transformer | |
| Tembusai et al. | K-nearest neighbor with K-fold cross validation and analytic hierarchy process on data classification | |
| CN109255389B (en) | Equipment evaluation method, device, equipment and readable storage medium | |
| CN115659966A (en) | Rumor detection method and system based on dynamic heteromorphic graph and multi-level attention | |
| Sandhu et al. | A comparative analysis of conjugate gradient algorithms & PSO based neural network approaches for reusability evaluation of procedure based software systems | |
| CN117290462B (en) | Intelligent decision system and method for large data model | |
| CN116522912B (en) | Training method, device, medium and equipment for package design language model | |
| CN113128689A (en) | Entity relationship path reasoning method and system for regulating knowledge graph | |
| CN116842263A (en) | Training processing method and device for intelligent question-answering financial advisor model | |
| CN111626844A (en) | Enterprise credit assessment method and device based on big data analysis | |
| CN112132219A (en) | General deployment scheme of deep learning detection model based on mobile terminal | |
| CN114091794A (en) | Patent value evaluation model training method, evaluation method, device and equipment | |
| CN113673680B (en) | Model verification method and system for automatically generating verification properties through an antagonism network | |
| CN113506186B (en) | Method and device for screening disturbed track of power system, electronic equipment and storage medium | |
| CN114820074A (en) | Target user group prediction model construction method based on machine learning | |
| CN114186646A (en) | Block chain abnormal transaction identification method and device, storage medium and electronic equipment | |
| CN111126617B (en) | Method, device and equipment for selecting fusion model weight parameters | |
| Kunwar | MWaste: A Deep Learning Approach to Manage Household Waste | |
| CN110766338A (en) | DPOS (distributed data processing) bifurcation prediction model method based on artificial intelligence and EOS (Ethernet over Ethernet) and IO (input/output) of block chain technology |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |