CN108595961A - The non-homogeneous software trust index fusion method matched based on ranking score - Google Patents
The non-homogeneous software trust index fusion method matched based on ranking score Download PDFInfo
- Publication number
- CN108595961A CN108595961A CN201810327392.0A CN201810327392A CN108595961A CN 108595961 A CN108595961 A CN 108595961A CN 201810327392 A CN201810327392 A CN 201810327392A CN 108595961 A CN108595961 A CN 108595961A
- Authority
- CN
- China
- Prior art keywords
- subinterval
- grade
- software
- trust index
- value
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
- G06F21/57—Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
- G06F21/577—Assessing vulnerabilities and evaluating computer system security
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/03—Indexing scheme relating to G06F21/50, monitoring users, programs or devices to maintain the integrity of platforms
- G06F2221/033—Test or assess software
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a kind of non-homogeneous software trust index fusion method matched based on ranking score, it can realize that the effective integration between software trust index calculates, and then complete the software credibility quantitative evaluation based on index, specifically include following steps:1) it is directed to and is distributed all kinds of software trust indexs heterogeneous, ratio of ratio and siding-to-siding block length shared by data in its dense distribution section and other sections etc. carries out ranking score and matches, and the relatively large number of grade of number is distributed for dense distribution section;2) ranking score based on software trust index interval is matched, and software trust index heterogeneous is converted to the number in [0,1] section, the normalization of non-homogeneous software trust index is realized, their fusion calculation is realized with weighted mean method.
Description
【Technical field】
The invention belongs to software trust index fusion calculation field, more particularly to it is a kind of matched based on ranking score it is non-homogeneous soft
Part credible indexes fusion method is realized not its object is to reduce the otherness between the software trust index of non-uniform Distribution
Same type, and the fusion calculation being distributed between software trust index heterogeneous.
【Background technology】
In software credibility quantitative evaluation, usually fusion calculation will be carried out after multiple software trust quantification of targets, such as
Weighted average, using obtained result of calculation as the quantitative evaluation value of software credibility.However, the value of software credibility index
Often there is type diversity and distribution characteristic heterogeneous, it is impossible to which fusion calculation is directly carried out to it.Such as extracting
Software creditability measurement parameter, be with " D`` PC5 " subset in NASA data sets MDP (Metrics Data Program)
Example, the value range of parameter CONDITION_COUNT is [0,1090], and the value range of parameter DESIGN_DENSITY is
The value of [0.0,1.0], parameter PARAMETER_COUNT is 0 to 90 natural number, they correspond to three kinds of different types respectively
Data.Meanwhile the value range of software metrics parameter NUM_OPERANDS is [0,5169], wherein 80% value concentrates on
In [0,122] this subinterval, and remaining 20% value be distributed in (122,5169] in this prodigious section, more especially
Larger value, value are even more sparse.These value types are various, software trust index unevenly distributed, it is difficult to be merged
It calculates.In order to realize the fusion calculation of software trust index various to value type and unevenly distributed, needing can to software
Believe that index carries out Homogeneous processing, and data normalization is a kind of common important method of data Homogeneous processing.Data normalizing
The big different types of data of difference can be converted to the real number in [0,1] section by change method, realize the unification of data format,
However traditional data normalization method can not improve the consistency between normalization data.For example, software metrics parameter
After NUM_OPERANDS normalization, 80% value concentrates in the subinterval of [0,0.024] this very little, and residue 20% takes
Distribution value (0.024,1] in this relatively large section.And after the value normalization of parameter PARAMETER_COUNT, phase
To being evenly distributed in [0,1] section, the number less than 0.1 is almost without such two software credibility indexs fusion meter
Obtained value will be without practical significance.
【Invention content】
The purpose of the present invention is to provide a kind of non-homogeneous software trust index fusion methods matched based on ranking score, in reality
Existing software index is normalized simultaneously, improves the consistency between software trust index, consequently facilitating carrying out software trust index
Fusion calculation, it is of the existing technology to solve the problems, such as.
In order to achieve the above objectives, the present invention is achieved by the following technical programs:
Based on the non-homogeneous software trust index fusion method that ranking score is matched, include the following steps:
1) acquisition software credible indexes value and its dense distribution section;Refer to for software credibility heterogeneous is distributed
Mark, the ratio of ratio and siding-to-siding block length shared by data in the dense distribution section of its value and other sections carry out etc.
Grade distribution;
2) ranking score of the distributed area based on software trust index value is matched, using the grade of distribution by non-homogeneous software
Credible indexes are converted to the real number in [0,1] section, realize the normalization of non-homogeneous software trust index.
The present invention further improvement lies in that, for software trust index heterogeneous is distributed in step 1), according to its value
Dense distribution section and other sections in the ratio of ratio and siding-to-siding block length shared by data carry out ranking score and match, need elder generation
Complete following operation:
First, software trust index value range is expressed as section [u, v], the section is intensive by data integrated distribution
Section [a, b] is divided into three subintervals:
PSR1:[u,a),PSR2:[a,b],PSR3:(b,v]
Wherein, [a, b] is data-intensive distributed area, PSRi(i=1,2,3) software trust index value range quilt is indicated
Three subintervals after division;
Secondly, it is each subinterval PSRi(i=1,2,3) calculates grade apportionment ratio, the purpose is to:1) it is subinterval point
With grade, the corresponding grade apportionment ratio in which subinterval being calculated is maximum, and next grade is assigned to corresponding sub-district
Between;2) prevent some subinterval from distributing too many grade, as subinterval ranking score matches the increase of number, ranking score is matched rate
It is gradually reduced, distributing to the number of levels in the subinterval will be reduced.The grade apportionment ratio in subinterval is calculated with following formula:
Wherein PiIndicate subinterval PSRiData rate of the data relative to total data in (i=1,2,3), wi(i=
1,2,3) siding-to-siding block length ratio of the siding-to-siding block length in subinterval relative to software trust index value range, a are indicatedi(i=1,2,
3) number of levels for being already allocated to subinterval is indicated,β is constant.In practical applications, β
It needs to be obtained according to the training of specific data set, in the present invention, according to the heterogeneity that software trust index value is distributed, if
Determine β=0.6.
The present invention further improvement lies in that, for software trust index heterogeneous is distributed in step 1), according to its value
Dense distribution section and other sections in the ratio of ratio and siding-to-siding block length shared by data carry out ranking score and match, it is specific to wrap
Include following steps:
A) according to wi(i=1,2,3) a is initializediIf wi=0, initialize ai=0, otherwise initialize ai=1, it is as long
Spend each the subinterval PSR divided being not zeroiDistribute an initial grade;
B) judging whether grade is assigned, that is, judge whether r≤g is true, wherein g indicates to want the number of allocation level,
It is set as 7, r in the present invention and indicates the number of levels being assigned;If r≤g is set up, goes to step c), otherwise terminate;
C) formula (1) is used to calculate the Q in each subintervali, i=1,2,3;
D) it calculatesSelect QiThe maximum subinterval of valueIt carries out ranking score to match, i.e.,It is right
It has been assigned number of levels and has added 1, i.e. r=r+1, return to step b).
The present invention further improvement lies in that, distributed area based on non-homogeneous software trust index value in step 2) etc.
Non-homogeneous software trust index is converted to the real number in [0,1] section using the grade of distribution, realizes software trust by grade distribution
The normalization of index, including the following contents:
First, each subinterval PSRi(i=1,2,3) the siding-to-siding block length h of corresponding grade ini(i=1,2,3) it calculates
It is as follows:
If ai=0, then correspond to hi=0, otherwise
Secondly, the value of the non-homogeneous software trust index to be converted is indicated with x, (k=1 2 ..., g) indicates that x is corresponded to k
Grade, αk(k=1,2 ..., g) indicate x be located at the ratio in section corresponding to grade k, k and αkIt is corresponding.It realizes non-homogeneous
The normalization of software trust index value needs to consider two kinds of situations, be respectively x and software the positive correlation of Credibility Assessment value with
It is negatively correlated.
If the Credibility Assessment value positive correlation of x and software, k and αkComputational methods it is as follows:
1) x belongs to subinterval PSR1=[u a) calculates the area that x subtracts grade in the lower bound u divided by the subinterval in subinterval
Between length h1, to obtained value round numbers part, the corresponding grade k of as x;With x subtract grade k corresponding under subinterval
The length h in subinterval corresponding to boundary divided by grade k1, obtain αk。
2) x belongs to subinterval PSR2=[a, b], the section of grade in the lower bound a divided by the subinterval in subinterval is subtracted with x
Length h2, to obtained value round numbers part, add section PSR1Interior number of levels a1, the corresponding grade k of as x;Use x
Subtract the length h in subinterval corresponding to the lower bound in subinterval divided by grade k corresponding to grade k2, obtain αk。
3) x belongs to subinterval PSR3=(b, v], the section of grade in the lower bound b divided by the subinterval in subinterval is subtracted with x
Length h3, to obtained value round numbers part, add section PSR1And PSR2Interior number of levels a1And a2, as x is corresponding
Grade k;With x subtract grade k corresponding to subinterval lower bound divided by grade k corresponding to subinterval length h3, obtain αk。
If the Credibility Assessment value of x and software is negatively correlated, the Credibility Assessment value first, in accordance with x and software is positively related
Method calculates k ' and α 'k, next k=g-k ', αk=1- α 'k。
The present invention further improvement lies in that, distributed area based on non-homogeneous software trust index value in step 2) etc.
Non-homogeneous software trust index is converted to the number in [0,1] section using the grade of distribution, realizes that software trust refers to by grade distribution
Target normalizes, and the fusion calculation of normalization software trust index, including the following contents are realized with weighted mean method:
First, the value x of software trust index is converted to the real number in [0,1] section with the following method:X is corresponded to first
Grade k be converted to real numberSecondly, by αkBe converted to real numberFinally calculateAs x it is corresponding [0,
1] real number in section.In the present invention, g values 7.
Secondly, x is usedi(i=1,2 ..., n) indicate software trust index, yi∈ [0,1] (i=1,2 ..., n) indicate it
Real number after normalization, then the metric for the software credibility that fusion calculation obtains is:
Wherein, ωi∈ [0,1] (i=1,2 ..., n) it is software trust index xiPower corresponding to (i=1,2 ..., n)
Weight.
Compared with the existing technology, the invention has the advantages that:The present invention is by can for the software of non-uniform Distribution
Believe that relatively large number of grade is distributed in the dense distribution section of index, is realizing that software trust index is normalized simultaneously, significantly carrying
Consistency between high software trust index, may be implemented that type is various and distribution software trust index heterogeneous is effectively melted
It is total to calculate.
【Description of the drawings】
Fig. 1 is the overall flow figure for the non-homogeneous software trust index fusion method matched the present invention is based on ranking score;
Fig. 2 is the flow chart that ratio of ratio and siding-to-siding block length shared by data etc. carries out that ranking score is matched.
【Specific implementation mode】
Below in conjunction with attached drawing, the present invention will be described in detail realizes the non-homogeneous software trust index fusion side matched based on ranking score
The embodiment of method.
Referring to Fig. 1, the present invention is based on the non-homogeneous software trust index fusion methods that ranking score is matched, and include the following steps:
Step S101:For software credibility index heterogeneous is distributed, according to the dense distribution section of its value and its
The ratio of ratio and siding-to-siding block length in its section shared by data carries out ranking score and matches;
Step S102:The ranking score of distributed area based on software trust index value is matched, will be non-using the grade of distribution
Uniform software trust index is converted to the real number in [0,1] section, realizes the normalization of non-homogeneous software trust index.
Specifically, for software trust index heterogeneous is distributed in step S101, according to the dense distribution of its value
The ratio of ratio and siding-to-siding block length in section and other sections shared by data carries out ranking score and matches, and needs first to complete following grasp
Make:
First, software trust index value range is expressed as section [u, v], the section is intensive by data integrated distribution
Section [a, b] is divided into three subintervals:
PSR1:[u,a),PSR2:[a,b],PSR3:(b,v]
Wherein, [a, b] is data-intensive distributed area, PSRi(i=1,2,3) software trust index value range quilt is indicated
Three subintervals after division;
Secondly, it is each subinterval PSRi(i=1,2,3) calculates grade apportionment ratio, the purpose is to:1) it is subinterval point
With grade, the corresponding grade apportionment ratio in which subinterval being calculated is maximum, and next grade is assigned to corresponding sub-district
Between;2) prevent some subinterval from distributing too many grade, as subinterval ranking score matches the increase of number, ranking score is matched rate
It is gradually reduced, distributing to the number of levels in the subinterval will be reduced.The grade apportionment ratio in subinterval is calculated with following formula:
Wherein PiIndicate subinterval PSRiData rate of the data relative to total data in (i=1,2,3), wi(i=
1,2,3) siding-to-siding block length ratio of the siding-to-siding block length in subinterval relative to software trust index value range, a are indicatedi(i=1,2,
3) number of levels for being already allocated to subinterval is indicated,β is constant.In practical applications, β
It needs to be obtained according to the training of specific data set, in the present invention, according to the heterogeneity that software trust index value is distributed, if
Determine β=0.6.
For example, in " D`` PC5 " subset of NASA data sets MDP (Metrics Data Program), software trust
The value of metric parameter HALSTED-LEVEL is distributed in section [0,2.0], wherein 84% value is distributed in intensive section
In [0,0.22], the section of the parameter value is divided into 2 subintervals:[0,0.22] and (0.22,2.0], then we can be with
It obtains, PSR1:[0,0],PSR2:[0,0.22],PSR3:(0.22,2.0], w1=0, w2=0.11, w3=0.89.According to statistics,
p1=0, p2=0.8311, p3=0.1699, initial setting, a1=0, a2=1, a3=1, β=0.6, then Qi(i=1,2,3)
It calculates as follows:
Referring to Fig. 2, specifically, for software trust index heterogeneous is distributed in step S101, according to its value
The ratio of ratio and siding-to-siding block length in dense distribution section and other sections shared by data carries out ranking score and matches, including as follows
Step:
S201 is according to wi(i=1,2,3) a is initializediIf wi=0, initialize ai=0, otherwise initialize ai=1, as
Each subinterval PSR divided that length is not zeroiDistribute an initial grade;
S202 judges whether grade is assigned, that is, judges whether r≤g is true, and wherein g indicates to want the number of allocation level
Mesh, general recommendations are set as 7, r and indicate the number of levels being assigned.If r≤g is set up, S203 is gone to step, is otherwise calculated
Method terminates;
S203 calculates the Q in each subinterval with formula (1)i(i=1,2,3);
S204 is calculatedSelect QiThe maximum subinterval of valueIt carries out ranking score to match, i.e.,Add 1, i.e. r=r+1, return to step S202 to being assigned number of levels.
For example, to software metrics parameter HALSTED-LEVEL, if relevant parameter is as previously mentioned, so its subinterval etc.
The process of grade distribution is as shown in the table:
Table 1 is the process of subinterval allocation level
Specifically, the ranking score of the distributed area based on non-homogeneous software trust index value in step 2) is matched, and is utilized
Non-homogeneous software trust index is converted to the real number in [0,1] section by the grade of distribution, realizes the normalizing of software trust index
Change, including the following contents:
First, each subinterval PSRi(i=1,2,3) the siding-to-siding block length h of corresponding grade ini(i=1,2,3) it calculates
It is as follows:
If ai=0, then correspond to hi=0, otherwise
Secondly, the value of the non-homogeneous software trust index to be converted is indicated with x, (k=1 2 ..., g) indicates that x is corresponded to k
Grade, αk(k=1,2 ..., g) indicate x be located at the ratio in section corresponding to grade k, k and αkIt is corresponding.It realizes non-homogeneous
The normalization of software trust index value needs to consider two kinds of situations, be respectively x and software the positive correlation of Credibility Assessment value with
It is negatively correlated.
If the Credibility Assessment value positive correlation of x and software, k and αkComputational methods it is as follows:
1) x belongs to subinterval PSR1=[u a) calculates the area that x subtracts grade in the lower bound u divided by the subinterval in subinterval
Between length h1, to obtained value round numbers part, the corresponding grade k of as x;With x subtract grade k corresponding under subinterval
The length h in subinterval corresponding to boundary divided by grade k1, obtain αk。
2) x belongs to subinterval PSR2=[a, b], the section of grade in the lower bound a divided by the subinterval in subinterval is subtracted with x
Length h2, to obtained value round numbers part, add section PSR1Interior number of levels a1, the corresponding grade k of as x;Use x
Subtract the length h in subinterval corresponding to the lower bound in subinterval divided by grade k corresponding to grade k2, obtain αk。
3) x belongs to subinterval PSR3=(b, v], the section of grade in the lower bound b divided by the subinterval in subinterval is subtracted with x
Length h3, to obtained value round numbers part, add section PSR1And PSR2Interior number of levels a1And a2, as x is corresponding
Grade k;With x subtract grade k corresponding to subinterval lower bound divided by grade k corresponding to subinterval length h3, obtain αk。
If the Credibility Assessment value of x and software is negatively correlated, the Credibility Assessment value first, in accordance with x and software is positively related
Method calculates k ' and α 'k, next k=g-k ', αk=1- α 'k。
Specifically, the ranking score of the distributed area based on non-homogeneous software trust index value in step 2) is matched, and is utilized
Non-homogeneous software trust index is converted to the number in [0,1] section by the grade of distribution, realizes the normalization of software trust index,
The fusion calculation of normalization software trust index, including the following contents are realized with weighted mean method:
First, the value x of software trust index is converted to the real number in [0,1] section with the following method:X is corresponded to first
Grade k be converted to real numberSecondly, by αkBe converted to real numberFinally calculateAs x it is corresponding [0,
1] real number in section.In the present invention, g values 7.
Secondly, x is usedi(i=1,2 ..., n) indicate software trust index, yi∈ [0,1] (i=1,2 ..., n) indicate it
Real number after normalization, then the metric for the software credibility that fusion calculation obtains is:
Wherein, ωi∈ [0,1] (i=1,2 ..., n) it is software trust index xiPower corresponding to (i=1,2 ..., n)
Weight.
For example, the value of software trust measurement parameter HALSTED-LEVEL is distributed in section [0,2.0], it can with software
Letter property quantitative evaluation value is positive correlation, and two values of the parameter are x1=0.2, x2=1.0.Software trust measures parameter
The value of PARAMETER_COUNT is 0 to 90 natural number, is negative correlativing relation, the ginseng with software credibility quantitative evaluation value
A several value x3=8.Obviously, x1,x2,x3Fusion calculation cannot directly be carried out.
In the preceding example, it is that the intensive section that HALSTED-LEVEL values are distributed (is expressed as PSR1) distribution grade
Number is a1=4, another section (is expressed as PSR2) distribution number of levels be a2=3, the number of levels to be distributed is 7, k1
And k2It is x1And x2Corresponding grade, μ (x1) and μ (x2) indicate by x1And x2The real number in section [0,1] being converted to, then
x1And x2It can be realized and be normalized by following process.
x1=0.2 ∈ PSR1, x2=1.0 ∈ PSR2, therefore have
Its integer part is 3, so k1=3
Its integer part is 1, so k2=1+4=5
The value of software trust measurement parameter PARAMETER_COUNT is to be uniformly distributed, x3=8 may be used traditional return
For the real number in [0,1] section, i.e., one changes method migrationu(x1),u(x2),u(x3) three values
Arithmetic weighted mean, is 0.4633.
Claims (5)
1. the non-homogeneous software trust index fusion method matched based on ranking score, which is characterized in that include the following steps:
1) acquisition software credible indexes value and its dense distribution section;For distribution software credibility index heterogeneous, root
Ranking score is carried out according to the ratio of ratio and siding-to-siding block length shared by data in the dense distribution section of its value and other sections
Match;
2) ranking score of the distributed area based on software trust index value is matched, using the grade of distribution by non-homogeneous software trust
Index is converted to the real number in [0,1] section, realizes the normalization of non-homogeneous software trust index, is returned with weighted mean method realization
One changes the fusion calculation of index.
2. the non-homogeneous software trust index fusion method according to claim 1 matched based on ranking score, which is characterized in that
For software trust index heterogeneous is distributed in step 1), according to data in the dense distribution section of its value and other sections
The ratio of shared ratio and siding-to-siding block length carries out ranking score and matches, and specifically includes:
First, software trust index value range is expressed as section [u, v], the section is by the intensive section of data integrated distribution
[a, b] is divided into three subintervals:
PSR1:[u,a),PSR2:[a,b],PSR3:(b,v]
Wherein, [a, b] is data-intensive distributed area, PSRiIndicate three sons after software trust index value range is divided
Section;I=1,2,3;
Secondly, it is each subinterval PSRiCalculate grade apportionment ratio;The grade apportionment ratio in subinterval is calculated with following formula:
Wherein PiIndicate subinterval PSRiData rate of the interior data relative to total data, wiIndicate that the section in subinterval is long
Spend the siding-to-siding block length ratio relative to software trust index value range, aiIndicate the number of levels for being already allocated to subinterval,β is constant.
3. the non-homogeneous software trust index fusion method according to claim 2 matched based on ranking score, which is characterized in that
For software trust index heterogeneous is distributed in step 1), according to data in the dense distribution section of its value and other sections
The ratio of shared ratio and siding-to-siding block length carries out ranking score and matches, and specifically comprises the following steps:
A) according to wiInitialize aiIf wi=0, initialize ai=0, otherwise initialize ai=1, as length is not zero each
The subinterval PSR of a divisioniDistribute an initial grade;
B) judge whether grade is assigned, that is, judge whether r≤g is true, wherein g indicates that the number of allocation level, r is wanted to indicate
The number of levels being assigned;If r≤g is set up, goes to step c), otherwise terminate;
C) formula (1) is used to calculate the Q in each subintervali, i=1,2,3;
D) it calculatesSelect QiThe maximum subinterval of valueIt carries out ranking score to match, i.e.,To
It is assigned number of levels and adds 1, i.e. r=r+1, return to step b).
4. the non-homogeneous software trust index fusion method according to claim 1 matched based on ranking score, which is characterized in that
The ranking score of distributed area based on non-homogeneous software trust index value in step 2) is matched, will be non-homogeneous using the grade of distribution
Software trust index is converted to the real number in [0,1] section, realizes the normalization of software trust index, including the following contents:
First, each subinterval PSRiThe siding-to-siding block length h of interior corresponding gradeiIt calculates as follows:
If ai=0, then correspond to hi=0, otherwise
Secondly, indicate the value of the non-homogeneous software trust index to be converted with x, k indicates the corresponding grades of x, k=1,2 ...,
G, αkIndicate that x is located at the ratio in section corresponding to grade k, k and αkIt is corresponding;Realize non-homogeneous software trust index value
Normalization considers two kinds of situations, is the positive correlation of Credibility Assessment value and the negative correlation of x and software respectively;
If the Credibility Assessment value positive correlation of x and software, k and αkComputational methods it is as follows:
1) x belongs to subinterval PSR1=[u, a), the section that calculating x subtracts grade in the lower bound u divided by the subinterval in subinterval are long
Spend h1, to obtained value round numbers part, the corresponding grade k of as x;With x subtract grade k corresponding to subinterval lower bound, remove
With the length h in subinterval corresponding to grade k1, obtain αk;
2) x belongs to subinterval PSR2=[a, b], the siding-to-siding block length of grade in the lower bound a divided by the subinterval in subinterval is subtracted with x
h2, to obtained value round numbers part, add section PSR1Interior number of levels a1, the corresponding grade k of as x;It is subtracted with x
The length h in subinterval corresponding to the lower bound in subinterval divided by grade k corresponding to grade k2, obtain αk;
3) x belongs to subinterval PSR3=(b, v], the siding-to-siding block length of grade in the lower bound b divided by the subinterval in subinterval is subtracted with x
h3, to obtained value round numbers part, add section PSR1And PSR2Interior number of levels a1And a2, the corresponding grades of as x
k;With x subtract grade k corresponding to subinterval lower bound divided by grade k corresponding to subinterval length h3, obtain αk;
If the Credibility Assessment value of x and software is negatively correlated, first, in accordance with the positively related method of the Credibility Assessment value of x and software
Calculate k ' and α 'k, next k=g-k ', αk=1- α 'k。
5. the non-homogeneous software trust index fusion method according to claim 1 matched based on ranking score, which is characterized in that
The ranking score of distributed area based on non-homogeneous software trust index value in step 2) is matched, will be non-homogeneous using the grade of distribution
Software trust index is converted to the number in [0,1] section, realizes the normalization of software trust index, is returned with weighted mean method realization
One changes the fusion calculation of index, including the following contents:
First, the value x of software trust index is converted to the real number in [0,1] section with the following method:It is first that x is corresponding etc.
Grade k is converted to real numberSecondly, by αkBe converted to real numberFinally calculateCorresponding [0,1] sections as x
Interior real number;
Secondly, x is usediIndicate software trust index, yi∈ [0,1] indicates the real number after its normalization, then what fusion calculation obtained
The metric of software credibility is:
Wherein, ωi∈ [0,1] is software trust index xiCorresponding weight;I=1,2 ..., n.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810327392.0A CN108595961B (en) | 2018-04-12 | 2018-04-12 | Non-uniform software credible index fusion method based on grade distribution |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810327392.0A CN108595961B (en) | 2018-04-12 | 2018-04-12 | Non-uniform software credible index fusion method based on grade distribution |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108595961A true CN108595961A (en) | 2018-09-28 |
CN108595961B CN108595961B (en) | 2021-07-02 |
Family
ID=63621943
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810327392.0A Active CN108595961B (en) | 2018-04-12 | 2018-04-12 | Non-uniform software credible index fusion method based on grade distribution |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108595961B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101763304A (en) * | 2009-12-31 | 2010-06-30 | 合肥工业大学 | Uncertain software credibility assessment method based on evidence theory |
CN102222040A (en) * | 2011-06-09 | 2011-10-19 | 西北工业大学 | Software creditability grade estimating method based on multiple-attribute entropy weight synthesis |
CN104102833A (en) * | 2014-07-10 | 2014-10-15 | 西安交通大学 | Intensive interval discovery based tax index normalization and fusion calculation method |
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
- 2018-04-12 CN CN201810327392.0A patent/CN108595961B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101763304A (en) * | 2009-12-31 | 2010-06-30 | 合肥工业大学 | Uncertain software credibility assessment method based on evidence theory |
CN102222040A (en) * | 2011-06-09 | 2011-10-19 | 西北工业大学 | Software creditability grade estimating method based on multiple-attribute entropy weight synthesis |
CN104102833A (en) * | 2014-07-10 | 2014-10-15 | 西安交通大学 | Intensive interval discovery based tax index normalization and fusion calculation method |
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 |
Also Published As
Publication number | Publication date |
---|---|
CN108595961B (en) | 2021-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109034244A (en) | Line loss abnormality diagnostic method and device based on electric quantity curve characteristic model | |
CN105488134A (en) | Big data processing method and big data processing device | |
CN106485529A (en) | The sort method of advertisement position and device | |
CN110134911A (en) | A kind of novel concatenation CSD constant multiplier algorithm structure for 256 fft processors | |
CN113315672B (en) | Flow distribution method and device, electronic equipment and computer readable storage medium | |
CN105139282A (en) | Power grid index data processing method, device and calculation device | |
CN107479856A (en) | Arctan function data structure and method for building up, function value-acquiring method and device | |
CN104102833A (en) | Intensive interval discovery based tax index normalization and fusion calculation method | |
DE102018126931A1 (en) | Apparatus and method based on direct anonymous attestation | |
CN108595961A (en) | The non-homogeneous software trust index fusion method matched based on ranking score | |
Liu et al. | Evaluation on regional science and technology resources allocation in China based on the zero sum gains data envelopment analysis | |
CN111538560B (en) | Virtual machine deployment method and device, electronic equipment and storage medium thereof | |
Arnold et al. | Unrestricted faithful rounding is good enough for some LNS applications | |
CN105763636B (en) | The selection method and system of optimal host in a kind of distributed system | |
CN108563504A (en) | A kind of resource management-control method and device | |
CN107392467A (en) | A kind of software work amount evaluation method and device based on unified demand | |
Gunter et al. | Multiple policy goals in a trade model with explicit factor markets | |
CN105120008B (en) | A kind of distributed cloud computing center balancing method of loads based on layering | |
Morris | Improved bounds for sampling contingency tables | |
Wang et al. | A nonlinear collective credit allocation in scientific publications | |
CN104093163B (en) | The business flow allocation method and device of terminal collaboration | |
CN110263073B (en) | Data value evaluation method and related device | |
CN107590105A (en) | Towards the computing device and method of nonlinear function | |
CN113283879A (en) | Method and device for determining construction amount, electronic equipment and readable storage medium | |
CN108650121B (en) | Comprehensive scoring and distributing method for multiple attribute Web service demand and provider |
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 |