CN105868551A - Fault association rule construction method - Google Patents

Fault association rule construction method Download PDF

Info

Publication number
CN105868551A
CN105868551A CN201610182316.6A CN201610182316A CN105868551A CN 105868551 A CN105868551 A CN 105868551A CN 201610182316 A CN201610182316 A CN 201610182316A CN 105868551 A CN105868551 A CN 105868551A
Authority
CN
China
Prior art keywords
frequent
item
candidate
items
item set
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201610182316.6A
Other languages
Chinese (zh)
Inventor
王艳辉
李曼
向万晓
贾利民
秦勇
林帅
王淑君
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.)
Beijing Jiaotong University
Original Assignee
Beijing Jiaotong 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 Beijing Jiaotong University filed Critical Beijing Jiaotong University
Priority to CN201610182316.6A priority Critical patent/CN105868551A/en
Publication of CN105868551A publication Critical patent/CN105868551A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
    • G16Z99/00—Subject matter not provided for in other main groups of this subclass

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a fault association rule construction method. The method comprises the following steps of 1, reading fault data records, recording a transaction code supporting each item and counting the number of transactions supporting each item, and obtaining a candidate k itemset Ck, wherein k=1; 2, adopting a tuning algorithm for the candidate k itemset for calculation to obtain a frequent k itemset Lk; 3, adopting a minimum candidate set generation method for connection pruning alternation on the frequent k itemset Lk to generate a candidate k+1 itemset Ck+1; 4, adopting the tuning algorithm for the candidate k+1 itemset for calculation to obtain a frequent k+1 itemset Lk+1; 5, judging whether the frequent k+1 itemset Lk+1 is the same as the frequent k itemset Lk or not; if yes, adopting the frequent k itemset Lk as the optimal frequent itemset; if not, k=k+1, and executing the step 3; 6, according to the optimal frequent itemset, constructing fault association rules. The method has the advantages of being high in stability, good in convergence behavior, high in running speed and the like.

Description

Fault association rule construction method
Technical Field
The invention relates to the technical field of fault association rule construction. And more particularly, to a fault association rule construction method.
Background
At present, research work at home and abroad mainly focuses on the identification of faulty elements. At present, the fault association rule construction methods include a logic processing method, an expert system method (an expert system method based on a Petri network), an artificial neuron network method, a method based on an optimization technology and the like. The optimization technology-based method comprises a genetic algorithm, a chaotic algorithm, a simulated annealing algorithm and the like. The expert system method in the methods is visual and has strong interpretation capability, but a complete knowledge base is difficult to obtain, no learning capability exists, and the fault-tolerant capability is poor. The performance of the fault association rule construction of the artificial neuron network method depends on whether a sample set is complete, and it is extremely difficult for a complex system to form a complete sample set, so that the correctness of the diagnosis result cannot be guaranteed in principle. The general fuzzy system adopts a structure similar to the expert system, so that the fuzzy system also has some inherent advantages and disadvantages of the expert system, but increases the fault tolerance. The Genetic Algorithm (GA) can basically solve the problem of fault association rule construction from the optimization point of view, and particularly can give a plurality of possible diagnosis results which are globally optimal or locally optimal under the conditions of complex faults or protection and circuit breaker misoperation, but has poor stability and low speed
Therefore, it is desirable to provide an accurate, efficient and stable fault association rule construction method.
Disclosure of Invention
The invention aims to provide an accurate, efficient and stable fault association rule construction method.
In order to achieve the purpose, the invention adopts the following technical scheme:
a fault association rule construction method comprises the following steps:
s1, reading the fault data record, recording the transaction code supporting each item and counting the number of the transactions supporting each item to obtain a candidate k item set Ck，k＝1；
S2 set C of candidate k itemskCalculating to obtain a frequent k item set L by adopting a tuning optimization algorithmk；
S3, set L of frequent k itemskCandidate k +1 item set C is generated by adopting a minimum candidate set generation method of connecting pruning for alternative operationk+1；
S4 set C of candidate k +1 itemsk+1Calculating to obtain a frequent k +1 item set L by adopting a tuning optimization algorithmk+1；
S5, judging frequent k +1 item set Lk+1Whether or not to match the frequent k term set LkThe same is that: if so, the frequent k item set LkAs the optimal frequent item set; if not, making k equal to k +1, and then the process proceeds to step S3;
and S6, constructing a fault association rule according to the optimal frequent item set.
Preferably, step S2 further includes the following sub-steps:
s2.1, calculating a candidate k item set CkThe support degree of (d) and the average value μ of the support degrees;
s2.2, counting to be larger than a minimum threshold value taukAnd calculating a preselection rate P of the preselected k-term setk；
S2.3, to the minimum threshold tau on condition of satisfying less than the standard deviation S of the support degreekAdjusting to select a preselected ratio PkThe optimal pre-selected k item set is used as the frequent k item set Lk。
Preferably, the minimum threshold τkTo maintain a candidate k term set CkThe minimum of support present.
Preferably, the support standard deviation S has a value formula as follows:
S = Σ i = 1 n ( X i - μ ) 2 / ( n - 1 )
wherein, XiThe support of the ith item in the pre-selected k item set is shown, and n is the number of items in the pre-selected k item set.
Preferably, the preselection rate P of the set of k items is preselectedkPreselection of k-term set number and candidate k-term set C for support above a minimum thresholdkThe ratio of the quantities.
Preferably, step S3 further includes the following sub-steps:
s3.1, set L of frequent k itemskThe first k-1 items are compared, and the frequent k item set L is deletedkDoes not have a sub-item set of the same front k-1 item, generates a frequent k item set L'k；
S3.2, calculating a frequent k item set LkEach item ini`Frequency of Lk(Ii`) I ═ 1,2,3 … n ', n' is a frequent k term set LkThe number of middle items, and the statistics of item sets I' with the frequency of the items less than k +2 ═ Ii|Lk(Ii)＜k+2}；
S3.3, deleting frequent k item set L'kThe frequent item set containing any item in the item set I' obtains a frequent k item set L ″k；
S3.4, set L' of frequent k itemskSelf-join generation of candidate k +1 term set Ck＋1。
Preferably, the set L' of frequent k itemskSelf-join generation of candidate k +1 term set Ck＋1The specific method is that if a ∈ L ″k,b∈L″kThen a + b ∈ Ck＋1A and b are respectively a frequent k item set L ″)kThe item of (1).
Preferably, step S3.4 further comprises the steps of: when k items are frequently collected, L ″)kIs empty, set L 'of frequent k items'kSelf-join generation of candidate k +1 term set Ck+1。
Preferably, the frequent k items are set L'kSelf-join generation of candidate k +1 term set Ck+1The concrete method is that if a ∈ L'k,b∈L′kThen a + b ∈ Ck＋1A and b are respectively frequent k term sets LkThe item of (1).
The invention has the following beneficial effects:
compared with the classical Apriori method, the technical scheme of the invention improves the efficiency by 23.75 percent and improves the accuracy by 7.39 percent. Simulation shows that the technical scheme of the invention not only can accurately construct the fault association rule, but also has good convergence rate and stability. Therefore, as a new application, the technical scheme of the invention has important theoretical significance and practical application value.
Drawings
The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.
FIG. 1 is a flow chart illustrating a method of fault association rule construction
Fig. 2 shows a flow chart of the tuning algorithm.
Fig. 3 shows a flow chart of a minimum candidate set generation method in which connection pruning is performed alternately.
Fig. 4 shows a schematic diagram of a raw fault data list.
Fig. 5 shows a schematic diagram of a candidate 1 item set list.
Fig. 6 shows a schematic diagram of a 1-item set tuning parameter list.
FIG. 7 shows a schematic diagram of a frequent 1 item set list.
Fig. 8 shows a schematic diagram of a candidate 2-item set list.
Fig. 9 shows a schematic diagram of a 2-item set tuning parameter list.
Fig. 10 shows a schematic diagram of a frequent 2-item set list.
Fig. 11 shows a schematic diagram of a candidate 3-item set list.
Fig. 12 shows a schematic diagram of a 3-item set tuning parameter list.
Fig. 13 shows a schematic diagram of a frequent 3 item set list.
Fig. 14 shows a schematic diagram of a candidate 4-item set list.
Fig. 15 shows a schematic diagram of a frequent 4-item set list.
Fig. 16 shows a schematic diagram of the constructed association rule table.
Detailed Description
In order to more clearly illustrate the invention, the invention is further described below with reference to preferred embodiments and the accompanying drawings. Similar parts in the figures are denoted by the same reference numerals. It is to be understood by persons skilled in the art that the following detailed description is illustrative and not restrictive, and is not to be taken as limiting the scope of the invention.
The construction of the fault association rule is to mine a Boolean association rule frequent item set to obtain occurrence relations among faults. The method for constructing the fault association rule provided by this embodiment changes a data structure from a horizontal structure to a vertical corresponding relationship data structure of a project transaction on the basis of a classic Apriori algorithm, records a transaction code supporting each item, extracts all items in the transaction, and counts the number of transactions supporting each item.
The connection idea in the minimum candidate set generation method in which the connection pruning is performed alternately in this embodiment is as follows: when generating a set of k +1 terms from a set of k terms, L is the set of frequent k termskWhen self-join is performed, if the first k-1 items of the two item sets are different, the join operation of the two item sets is abandoned, because the generated item sets are not repeated or are not frequent item sets. I.e. setting1And l2Is LkIn ascending order by the frequency with which items in the set of items appear throughout the database, performing Lk×LkWhen, if l1(k-1)≠l2(k-1)Then l may be discarded1And l2Otherwise, the generated k +1 item set must be a redundant item set, so that the calculation amount can be reduced.
The pruning idea in the minimum candidate set generation method in which the connection pruning is performed alternately in this embodiment is as follows: if the set of k terms X ═ X1,x2,…,xkIn (f), there is one xi∈ X makes Lk-1(Ii) < k-1, then X is not a frequent item set, where IiRepresenting a frequent k-1 term set Lk-1Contains x in the set ofiThe number of (2).
As shown in fig. 1, the method comprises the steps of:
s1, reading the fault data record, recording the transaction code supporting each item and counting the number of the transactions supporting each item to obtain a candidate k item set CkK is 1, thus saving the time wasted in pattern matching the candidate transaction;
s2 set C of candidate k itemskCalculating to obtain a frequent k item set L by adopting a tuning optimization algorithmk；
S3, set L of frequent k itemskCandidate k +1 item set C is generated by adopting a minimum candidate set generation method of connecting pruning for alternative operationk+1；
S4 set C of candidate k +1 itemsk＋1Calculating to obtain a frequent k +1 item set L by adopting a tuning optimization algorithmk＋1；
S5, judging frequent k +1 item set Lk＋1Whether or not to match the frequent k term set LkThe same is that: if so, the frequent k item set LkAs the optimal frequent item set; if not, making k equal to k +1, and then the process proceeds to step S3;
and S6, constructing a fault association rule according to the optimal frequent item set.
Wherein,
as shown in fig. 2, step S2 further includes the following sub-steps:
s2.1, calculating a candidate k item set CkThe support degree of (d) and the average value μ of the support degrees;
s2.2, counting to be larger than a minimum threshold value taukAnd calculating a preselection rate P of the preselected k-term setk；
S2.3, to the minimum threshold tau on condition of satisfying less than the standard deviation S of the support degreekMake an adjustmentSelecting a preselected ratio PkThe optimal pre-selected k item set is used as the frequent k item set LkI.e. the standard deviation of the support in terms of the mean value of the supports mu, neighborhood of S, i.e. [ mu + S, [ mu ] -S]In a moderate range and at the same time as a minimum threshold interval, the minimum threshold taukThe adjustment is within the minimum threshold interval [ mu + S, mu-S]Internally valued to find the optimal frequent k term set Lk。
Minimum threshold τkTo maintain a candidate k term set CkThe minimum of the existing support, the final minimum threshold τ for all sets of terms in the same layer (e.g., k-layer and k-layer) is equal, while the minimum threshold τ for two different layers (e.g., k-layer and k + 1-layer) is unequal.
The value formula of the standard deviation S of the support degree is as follows:
S = &Sigma; i = 1 n ( X i - &mu; ) 2 / ( n - 1 )
wherein, XiThe support of the ith item in the pre-selected k item set is shown, and n is the number of items in the pre-selected k item set.
Preselection rate P of preselection k item setkPreselection of k-term set number and candidate k-term set C for support above a minimum thresholdkThe ratio of the quantities.
As shown in fig. 3, step S3 further includes the following sub-steps:
s3.1, set L of frequent k itemskThe first k-1 items are compared, and the frequent k item set L is deletedkDoes not have a sub-item set of the same front k-1 item, generates a frequent k item set L'k；
S3.2, calculating a frequent k item set LkEach item ini`Frequency of Lk(Ii`) I ═ 1,2,3 … n ', n' is a frequent k term set LkThe number of middle terms, and the statistics of term set I' with frequency less than k +2 ═ Ii|Lk(Ii)＜k+2}，
S3.3, deleting frequent k item set L'kThe frequent item set containing any item in the item set I' obtains a frequent k item set L ″k；
S3.4, set L' of frequent k itemskSelf-join generation of candidate k +1 term set Ck+1。
In step S3.4, the frequent k item set L ″)kSelf-join generation of candidate k +1 term set Ck+1The specific method is that if a ∈ L ″k,b∈L″kThen a + b ∈ Ck＋1I.e. calculating Ta＋b＝Ta∪TbA and b are respectively a frequent k item set L ″)kItem of (1), Ta、TbRespectively a set L of frequent k itemskThe sub-item set containing a and b, Ta＋bSet L "of frequent k itemskContains both a and b sub-item sets.
Step S3.4 also includes the steps of: when k items are frequently collected, L ″)kIs empty, set L 'of frequent k items'kSelf-join generation of candidate k +1 term set Ck＋1。
And L 'will be set of frequent k terms'kSelf-join generation of candidate k +1 term set Ck＋1The concrete method is that if a ∈ L'k,b∈L′kThen ab ∈ Ck＋1I.e. calculating Ta＋b＝Ta∪TbA and b are respectively set L 'of frequent k items'kItem of (1), Ta、TbRespectively set of frequent k terms L'kThe sub-item set containing a and b, Ta＋bIs a set of frequent k terms L'kContains both a and b sub-item sets.
The method for constructing the fault association rule provided in this embodiment is further described below by substituting specific fault data.
There are 68 pieces of fault data, and there are 26 kinds of fault attributes (relevant parameters and components of fault, etc.), as shown in fig. 4. The basic steps of the method for constructing the fault association rule provided by this embodiment are as follows:
1) changing a data structure, recording a transaction code supporting each item, and counting the number of transactions supporting each item to obtain a candidate 1 item set, wherein specific data are shown in fig. 5;
2) adopting tuning optimization algorithm to carry out on candidate 1 item set C1Calculating to obtain a frequent 1 item set L1：
First, a candidate 1 item set C is calculated1The support degree of each attribute in (1) is calculated, and the average value μ and the standard error S are calculated to obtain a minimum threshold interval [31,39.6 ] by setting μ to 35.3 and S to 4.3]. Taking a minimum threshold τ1Equal to the average of 35.3, then preselection rate of 1 item setMinimum threshold τ1In the interval [31,39.6]Randomly taking values for several times, calculating preselection rate, and comparing preselection rate P1Selecting P1Minimum threshold σ closer to 0.51To obtain a better frequent 1 item set L1The details are shown in fig. 6 and 7.
3) Frequent 1 item set L1Performing self-connection operation to obtain a candidate 2 item set C2；
Specific data As shown in FIG. 8, since the candidate 2 items set C2The resulting data is so much that only the first part is intercepted.
4) Adopting tuning optimization algorithm to carry out on candidate 2 item set C2Calculating to obtain a frequent 2 item set L2：
First, a candidate 2 item set C is calculated2Calculating the support of each attribute set, and obtaining the minimum threshold interval [18.1,24.1 ] by calculating the average value mu and the standard error S, and obtaining the minimum threshold interval [ 21.1 and 3.0 ] by the value mu and the standard error S]. Taking a minimum threshold τ2Equal to an average value of 21.1, thenPreselection of 2 item set preselection ratesMinimum threshold τ2In the interval [18.1,24.1]Randomly taking values for several times, calculating preselection rate, and comparing preselection rate P2Selecting P2Minimum threshold τ closer to 0.52To obtain a better frequent 2 item set L2The details are shown in FIGS. 9 and 10.
5) The method for generating the minimum candidate set by alternately performing the connection pruning is adopted to carry out the L on the frequent 2 item sets2Calculating to obtain a candidate 3 item set C3：
Comparing the first 1 items of the attribute set, and deleting L2Does not have attribute set of the same top 1 item, generates a new frequent 2 item set L'2Then calculate L2Each item iniFrequency of (D), marked as L2(Ii) Finding out the item with the item frequency less than 4, and marking as I', I ═ Ii|L2(Ii) < 4 >, scan L'2L 'is removed'2The frequent item set containing any item in I' obtains a new smaller frequent 2 item set L ″2Then L ″'2Generating candidate 3-item set C from joins3The specific data are shown in fig. 9.
6) Adopting tuning optimization algorithm to carry out optimization on candidate 3 item set C3Calculating to obtain a frequent 3 item set L3：
First, a candidate 3 item set C is calculated3Calculating the support of each attribute set, and obtaining the minimum threshold interval [16.6,19.8 ] by calculating the average value mu and the standard error S, and obtaining the minimum threshold interval [ 18.2 ] and the minimum threshold interval S as 1.6]. Taking a minimum threshold τ3Equal to an average of 18.2, a preselection rate of 3 sets of items is preselectedMinimum threshold τ3In the interval [16.6,19.8 ]]Randomly taking values for several times, calculating preselection rate, and comparing preselection rate P3Selecting P3Minimum threshold τ closer to 0.53To obtain betterFrequent 3 item set L3The details are shown in FIGS. 12 and 13.
7) The method for generating the minimum candidate set by alternately performing the connection pruning is adopted to carry out on the frequent 3 item sets L3Calculating to obtain a candidate 4 item set C4：
Comparing the first 2 items of the attribute set, and deleting L2Does not have attribute set of the same top 2 items, generates a new frequent 3 item set L'3Then calculate L3Each attribute I iniFrequency of (D), marked as L3(Ii) Find out the item with frequency less than 5, and mark as I', I ═ Ii|L3(Ii) < 5 >, scan L'3L 'is removed'3The frequent item set containing any item in I' obtains a new smaller frequent 3 item set L ″3. Due to L ″)3Is empty, make L'3Generating candidate 4-item set C from joins4The details are shown in FIG. 14.
8) At this time, only three attribute sets { A, G, K, Q }, { K, L, N, Q }, { K, N, Q }, and { K, N, Q, W } with the support degrees of 15, 14, and 15 remain, the attribute set { K, L, N, Q } with the support degree of 14 can be directly deleted, and a frequent 4-item set L is obtained4The concrete data are shown as 15;
9) due to frequent 4 item set L4There is no top 3 identical property set, L'4If the set is empty, the join operation cannot be performed, and the set L of 4 items is frequent4The optimal frequent item set is { { a, G, K, Q }, { K, N, Q, W } }.
Thus, two fault association rules can be derived, and substituting letters for meaning can yield a fault association rule as shown in fig. 16.
It should be understood that the above-mentioned embodiments of the present invention are only examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention, and it will be obvious to those skilled in the art that other variations or modifications may be made on the basis of the above description, and all embodiments may not be exhaustive, and all obvious variations or modifications may be included within the scope of the present invention.

Claims (9)

1. A fault association rule construction method is characterized by comprising the following steps:
s1, reading the fault data record, recording the transaction code supporting each item and counting the number of the transactions supporting each item to obtain a candidate k item set Ck，k＝1；
S2 set C of candidate k itemskCalculating to obtain a frequent k item set L by adopting a tuning optimization algorithmk；
S3, set L of frequent k itemskCandidate k +1 items are generated by adopting a minimum candidate set generation method of connecting pruning in an alternating mannerCollection Ck+1；
S4 set C of candidate k +1 itemsk+1Calculating to obtain a frequent k +1 item set L by adopting a tuning optimization algorithmk+1；
S5, judging frequent k +1 item set Lk+1Whether or not to match the frequent k term set LkThe same is that: if so, the frequent k item set LkAs the optimal frequent item set; if not, making k equal to k +1, and then the process proceeds to step S3;
and S6, constructing a fault association rule according to the optimal frequent item set.
2. The method of claim 1, wherein the step S2 further comprises the following sub-steps:
s2.1, calculating a candidate k item set CkThe support degree of (d) and the average value μ of the support degrees;
s2.2, counting to be larger than a minimum threshold value taukAnd calculating a preselection rate P of the preselected k-term setk；
S2.3, to the minimum threshold tau on condition of satisfying less than the standard deviation S of the support degreekAdjusting to select a preselected ratio PkThe optimal pre-selected k item set is used as the frequent k item set Lk。
3. The method of claim 2, wherein the minimum threshold τ is set tokTo maintain a candidate k term set CkThe minimum of support present.
4. The method for constructing the fault association rule according to claim 2, wherein a numeric formula of the standard deviation of the support degree S is as follows:
S = &Sigma; i = 1 n ( X i - &mu; ) 2 / ( n - 1 )
wherein, XiThe support of the ith item in the pre-selected k item set is shown, and n is the number of items in the pre-selected k item set.
5. The method of constructing a fault association rule according to claim 2, characterized in that the preselection rate P of the k term set is preselectedkPreselection of k-term set number and candidate k-term set C for support above a minimum thresholdkThe ratio of the quantities.
6. The method of claim 1, wherein the step S3 further comprises the following sub-steps:
s3.1, set L of frequent k itemskThe first k-1 items are compared, and the frequent k item set L is deletedkDoes not have a sub-item set of the same front k-1 item, generates a frequent k item set L'k；
S3.2, calculating a frequent k item set LkEach item ini`Frequency of Lk(Ii`) I ═ 1,2,3 … n ', n' is a frequent k term set LkThe number of middle items, and the statistics of item sets I' with the frequency of the items less than k +2 ═ Ii|Lk(Ii)＜k+2}；
S3.3, deleting frequent k item set L'kThe frequent item set containing any item in the item set I' obtains a frequent k item set L ″k；
S3.4, set L' of frequent k itemskSelf-join generation of candidate k +1 term set Ck+1。
7. The method according to claim 6, wherein the set of frequent k items L ″, is defined askSelf-join generation of candidate k +1 term set Ck+1The specific method is that if a ∈ L ″k,b∈L″kThen a + b ∈ Ck+1A and b are respectively a frequent k item set L ″)kThe item of (1).
8. The method of claim 6, wherein step S3.4 further comprises the steps of: when k items are frequently collected, L ″)kIs empty, set L 'of frequent k items'kSelf-join generation of candidate k +1 term set Ck+1。
9. The fault association rule construction method according to claim 6, wherein the set of frequent k items L'kSelf-join generation of candidate k +1 term set Ck+1The concrete method is that if a ∈ L'k,b∈L′kThen a + b ∈ Ck+1A and b are respectively set L 'of frequent k items'kThe item of (1).
CN201610182316.6A 2016-03-28 2016-03-28 Fault association rule construction method Pending CN105868551A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610182316.6A CN105868551A (en) 2016-03-28 2016-03-28 Fault association rule construction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610182316.6A CN105868551A (en) 2016-03-28 2016-03-28 Fault association rule construction method

Publications (1)

Publication Number Publication Date
CN105868551A true CN105868551A (en) 2016-08-17

Family

ID=56626092

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610182316.6A Pending CN105868551A (en) 2016-03-28 2016-03-28 Fault association rule construction method

Country Status (1)

Country Link
CN (1) CN105868551A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106779305A (en) * 2016-11-22 2017-05-31 南方电网科学研究院有限责任公司 Customer interaction trace analysis method and device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819411A (en) * 2010-03-17 2010-09-01 燕山大学 GPU-based equipment fault early-warning and diagnosis method for improving weighted association rules
CN102809965A (en) * 2012-07-30 2012-12-05 燕山大学 Fault early warning method for hydraulic equipment based on fault frequent pattern
CN103728507A (en) * 2013-11-18 2014-04-16 芜湖大学科技园发展有限公司 Grid fault diagnosis method based on data mining
CN103793589A (en) * 2012-10-31 2014-05-14 中国科学院软件研究所 High-speed train fault handling method
CN104268375A (en) * 2014-09-10 2015-01-07 中国人民解放军海军工程大学 Ship electric power station fault diagnosing method based on knowledge petri network

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819411A (en) * 2010-03-17 2010-09-01 燕山大学 GPU-based equipment fault early-warning and diagnosis method for improving weighted association rules
CN102809965A (en) * 2012-07-30 2012-12-05 燕山大学 Fault early warning method for hydraulic equipment based on fault frequent pattern
CN103793589A (en) * 2012-10-31 2014-05-14 中国科学院软件研究所 High-speed train fault handling method
CN103728507A (en) * 2013-11-18 2014-04-16 芜湖大学科技园发展有限公司 Grid fault diagnosis method based on data mining
CN104268375A (en) * 2014-09-10 2015-01-07 中国人民解放军海军工程大学 Ship electric power station fault diagnosing method based on knowledge petri network

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
李曼: ""基于机器学习的故障识别方法与系统研制"", 《中国优秀硕士论文全文数据库 工程科技Ⅱ辑》 *
章句才: "《工业噪声测量指南》", 31 March 1984, 北京计量出版社 *
蔡红 等: "关联规则挖掘最小支持度阀值设定的优化算法研究"", 《微型电脑应用》 *
陈波: ""一种关联规则挖掘最小支持度调优算法"", 《科技创新导报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106779305A (en) * 2016-11-22 2017-05-31 南方电网科学研究院有限责任公司 Customer interaction trace analysis method and device
CN106779305B (en) * 2016-11-22 2021-05-14 南方电网科学研究院有限责任公司 Customer interaction trace analysis method and device

Similar Documents

Publication Publication Date Title
Dunne et al. Firm entry and postentry performance in the US chemical industries
CN104462184B (en) A kind of large-scale data abnormality recognition method based on two-way sampling combination
CN103927346B (en) Query connection method on basis of data volumes
TW201237647A (en) Method and system for identifying rare-event failure rates
CN104537067A (en) Box separation method based on k-means clustering
CN106548418B (en) Small interference stability evaluation method for power system
CN109543765A (en) A kind of industrial data denoising method based on improvement IForest
CN104598569A (en) Association rule-based MBD (Model Based Definition) data set completeness checking method
CN111638707B (en) Intermittent process fault monitoring method based on SOM clustering and MPCA
Parssian Managerial decision support with knowledge of accuracy and completeness of the relational aggregate functions
CN111401573A (en) Working condition state modeling and model correcting method
CN115270007B (en) POI recommendation method and system based on mixed graph neural network
CN113722997A (en) New well dynamic yield prediction method based on static oil and gas field data
CN113641733B (en) Real-time intelligent estimation method for river cross section flow
CN113886183B (en) Method for measuring and calculating occurrence time of voltage sag event
CN114118730A (en) Compliance detection method for business process of power enterprise
CN105868551A (en) Fault association rule construction method
CN109635008B (en) Equipment fault detection method based on machine learning
CN115713270B (en) Method and device for detecting and correcting peer mutual evaluation abnormal scores
Wang et al. Stacking based LightGBM-CatBoost-RandomForest algorithm and its application in big data modeling
CN104036024B (en) It is a kind of based on GACUC and the spatial clustering method of Delaunay triangulation network
CN112115620B (en) Learning curve product unit cost evaluation method based on fractal thought
Ko et al. A multivariate parameter trace analysis for online fault detection in a semiconductor etch tool
Last et al. Data mining for process and quality control in the semiconductor industry
Fewster et al. Similarity indices for spatia I ecological data

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20160817