CN105787610A - Case-based reasoning method capable of supporting time sequence matching - Google Patents
Case-based reasoning method capable of supporting time sequence matching Download PDFInfo
- Publication number
- CN105787610A CN105787610A CN201410797922.XA CN201410797922A CN105787610A CN 105787610 A CN105787610 A CN 105787610A CN 201410797922 A CN201410797922 A CN 201410797922A CN 105787610 A CN105787610 A CN 105787610A
- Authority
- CN
- China
- Prior art keywords
- case
- dist
- characteristic item
- distance
- reasoning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a case-based reasoning method capable of supporting time sequence matching. The method includes the following steps of: case base construction; case retrieval: with the state information of a current target event adopted as input, case retrieval is carried out based on the case base, so that the comprehensive distance from the current target event to cases in the case base can be obtained; case reuse: if the comprehensive distance from the current target event to a certain case in the case base is smaller than a set threshold value, the current target event is successfully matched with the case, and the case result of the case is outputted so as to be used for a target event; case correction: when the case is wrong, the successfully-matched case result is corrected according to the feedback information of the target event, and the corrected case result is submitted to the case base, and case-based reasoning is terminated. According to the case-based reasoning method capable of supporting time sequence matching of the invention, a time sequence matching mechanism is introduced into case feature item matching, so that case-based reasoning has a time trend analysis ability, and can be better applied to the engineering field emphasizing time trend analysis.
Description
Technical field
The present invention relates to a kind of reasoning by cases method, particularly a kind of reasoning by cases method supporting Time Series Matching.The method belongs to artificial intelligence field.
Background technology
Reasoning by cases (case-basedreasoning, CBR) is an emerging field of artificial intelligence.The problem solving method of case-based reasioning, be highly suitable for not very strong theoretical model and domain knowledge not exclusively, be difficult to define or define in inconsistent and veteran policy setting, be all widely used in fields such as medical diagnosis, legal advice, project planning and fault diagnosises at present.
At present, the core link of reasoning by cases is Case Retrieval, namely find in case library and describe the most close case with problem, it depends on the coupling of case characteristic item, but case characteristic item is mostly towards single numerical value at present, being not carried out seasonal effect in time series support, this makes reasoning by cases relatively limited in some application laying particular stress on time trend coupling.
Summary of the invention
For above-mentioned technical deficiency, the present invention proposes a kind of reasoning by cases method supporting Time Series Matching, its objective is: Time Series Matching mechanism is incorporated in reasoning by cases, increase case characteristic item to seasonal effect in time series support, and solve seasonal effect in time series matching distance based on dynamic time warping distance method (Dynamictimewarping, DTW)；By adopting the case characteristic item weight adjustment algorithm with matching distance proportion coefficients to realize the study correction to Feature item weighting, make reasoning by cases more engineering practicability.
The technical solution adopted for the present invention to solve the technical problems is: a kind of reasoning by cases method supporting Time Series Matching, comprises the following steps:
1) structure case library；
2) Case Retrieval: with the status information of current goal event for input, Design case based storehouse carries out Case Retrieval, obtains current goal event and the comprehensive distance of each case in case library；
3) case is reused: if current goal event is with case library, the comprehensive distance of certain case is less than setpoint distance threshold value, then the match is successful with this case for current goal event, the case result of this case is exported and is used for object event, performs next step；Otherwise, it fails to match and terminates；
4) Case-based adaptation: judge that whether the case result that the match is successful is consistent with current goal event；If consistent, then this case is correct, and reasoning by cases terminates；Otherwise this case mistake, is modified the case result that the match is successful according to the feedback information of object event, and revised case result is submitted to case library, and reasoning by cases terminates.
Described case library is the set of multiple case, and described case includes: case title, case characteristic item set, case result, case effect assessment.
Described Case Retrieval formula is as follows:
In above formula, SimkRepresent kth case X in case librarykCharacteristic item sequence X with current goal event0Comprehensive distance, ωkjFor the weight that the jth attribute of kth case is shared in participating in the ATTRIBUTE INDEX of case coupling, j=1,2 ..., m, m is the attribute number of kth case；Dist(X0(j),Xk(j)) represent the characteristic item sequence X of kth case and current goal event0Matching distance on jth attribute.
The characteristic item of described object event is time point value, then Dist (X on jth attribute0(j),Xk(j)) for manhatton distance | X0(j),Xk(j)|；X0The j characteristic item sequence that () is current goal event time point value on jth attribute, XkJ () is kth case time point value on jth attribute.
The characteristic item of described object event is time series on jth attribute, Dist (X0(j),Xk(j)) in dynamic time warping distance algorithm consolidation path distance D (| Xk(j)|,|X0(j)|)；X0The j characteristic item sequence that () is current goal event time series on jth attribute, XkJ () is kth case time series on jth attribute.
A kind of reasoning by cases method supporting Time Series Matching, also includes the weight of case characteristic item is adjusted, and adjusts amplitude formula by weight and realizes:
W'kj=Wkj(1±Δ×Rkj)(2)
W in above formulakjRepresent characteristic item j weight before correction in case k, W'kjRepresenting that in case k, characteristic item j is in revised weight, Δ represents the range coefficient of adjustment every time, RkjRepresenting the matching distance proportion coefficients of characteristic item j in case k, computing formula is as follows:
In above formula, SimkRepresent kth case X in case librarykCharacteristic item sequence X with current goal event0Comprehensive distance, TDistRepresent characteristic item matching distance threshold value, Dist (X0(j),Xk(j)) represent the characteristic item sequence X of kth case and current goal event0Matching distance on jth attribute；ωkjFor the weight that the jth attribute of kth case is shared in the ATTRIBUTE INDEX participating in case coupling.
When case k is correct and Dist (X0(j),Xk(j)/Simk>=TDistTime:
W'kj=Wkj(1-Δ×ωkjDist(X0(j),Xk(j)/Simk)。
When case k is correct and Dist (X0(j),Xk(j)/Simk< TDistTime:
W'kj=Wkj(1+Δ×(1-ωkjDist(X0(j),Xk(j)/Simk))。
When case k is judged as mistake and Dist (X0(j),Xk(j)/Simk>=TDistTime:
W'kj=Wkj(1+Δ×ωkjDist(X0(j),Xk(j)/Simk)。
When case k is judged as mistake and Dist (X0(j),Xk(j)/Simk< TDistTime:
W'kj=Wkj(1-Δ×(1-ωkjDist(X0(j),Xk(j)/Simk))。
The invention have the advantages that and advantage:
1. Time Series Matching mechanism is incorporated in case characteristic item coupling by the present invention, makes reasoning by cases have time trend analysis ability, it is possible to be preferably applied for laying particular stress on the engineering field of time trend analysis.
2. in the expression structure of case, adding the case effect assessment reusing effect for recording case, it is possible to being easy in routine use constantly to add evaluation effect is excellent case, deleting evaluation effect is bad case, improve case library so that it is more engineering practicability.
3. the support manhatton distance adopted in Case Retrieval and the nearest neighbor method of dynamic time warping distance, case characteristic item matching distance can be solved from time point value and time series the two aspect, the comprehensive distance finally drawn more can reflect the similarity of case and current goal event, and improves the accuracy of Case Retrieval.
4. the Feature item weighting adjustment algorithm adopted can reuse situation by Design case based, revises case Feature item weighting, and then makes case characteristic item coupling in the future more conform to practical situation, has important practical significance and higher engineer applied is worth.
Accompanying drawing explanation
Fig. 1 is the flow chart of a kind of reasoning by cases method supporting Time Series Matching of the present invention.
Fig. 2 is the flow chart that case characteristic item weight adjusts.
Detailed description of the invention
Below in conjunction with embodiment, the present invention is described in further detail.
In order to make the purpose of the present invention, technical scheme and advantage clearly understand, below in conjunction with embodiment, the present invention is described in further detail.Should be appreciated that specific embodiment described herein is only in order to explain the present invention, is not intended to limit the present invention.
Present invention can apply to manufacture field, the reasoning by cases method of the support Time Series Matching present invention proposed initially with programming language carries out code realization, and it is packaged into software, then carrying out software installation with the PC of industry spot, server or industrial computer for carrier, last Field Force can carry out the function such as process control, accident analysis according to the reasoning by cases result that software shows.
Specific design step is as follows:
Step one: structure case library, method particularly includes:
The case library of the present invention is the set of multiple cases that manufacture field was collected, and the expression structure of case specifically includes: case title, the characteristic item set of case, case result, case effect assessment；Wherein, case characteristic item set, case effect assessment and case result are the core component of case.
(1) characteristic item in the set of case characteristic item mainly describes in manufacturing industry scene, for instance the numerical characteristics of the monitoring parameter such as electric current, voltage, it both can be time point value, it is also possible to be time series.Time point value describes the jump feature of monitoring parameter, mainly through catching the monitoring numerical value of correspondence monitoring parameter when saltus step occurs, obtains the time point value of characteristic item；Time series describes the Long-term change trend feature of monitoring parameter, in order to save the memory space of case, here initial time and the deadline of corresponding time trend are only stored, when the case characteristic item carrying out time series type mates, it is possible to directly extract complete time series from data base according to initial time and deadline.
(2) case effect assessment is based on spot effect after case exports manufacturing industry on-site target event and carries out adding typing, manufacturing industry on-site target event refers mainly to the rational analysis carried out for field apparatus or operation process, whether such as analyze the operation stability of production line, to analyze pump capacity normal etc., and case effect assessment is divided into Three Estate excellent, good, bad.If case result is consistent with current goal event, then the effect assessment of corresponding case is updated to excellent；Otherwise effect assessment is updated to bad.For case library, it is simply that will constantly add evaluation effect is excellent case, deleting evaluation effect is bad case, and then forms comparatively complete case library.
(3) case result refers under corresponding monitoring parameter, the process operation result that manufacturing industry scene occurs, for instance motor speed exception, production line fault etc..A case for following: very low at rotating speed, under the state that shell temperature is significantly high, motor burns out.Then in this case, " motor burns out " is case result.
Step 2: Case Retrieval, method particularly includes:
With the monitoring parameter information of manufacturing industry scene current goal event for input, Design case based storehouse carries out Case Retrieval, obtains current goal event and the comprehensive distance of each case in case library.
Time trend coupling has been dissolved in nearest neighbor method by the case retrieving method of present invention design, supports the manhatton distance between attribute and dynamic time warping distance, and case retrieval algorithm formula is as follows:
In above formula, SimkRepresent kth case X in case librarykCharacteristic item sequence X with current goal event0Comprehensive distance, case XkWith characteristic item sequence X0Form by a series of field monitoring parameter values, ωkjFor the weight that the jth attribute of kth case is shared in participating in the ATTRIBUTE INDEX of case coupling, j=1,2 ..., m, m is the attribute number of kth case；Dist(X0(j),Xk(j)) represent the characteristic item sequence X of kth case and current goal event0Matching distance on jth attribute, wherein, jth attribute i.e. jth field monitoring parameter.
Characteristic item sequence X0If jth field monitoring parameter be time point value, Dist (X0(j),Xk(j)) it is manhatton distance | X0(j),Xk(j)|；XkJ () is kth case time point value on jth attribute, X0The j characteristic item sequence that () is current goal event time point value on jth attribute.
Characteristic item sequence X0If jth field monitoring parameter be seasonal effect in time series words, Dist (X0(j),Xk(j)) be consolidation path distance D in dynamic time warping distance algorithm (| Xk(j)|,|X0(j) |), wherein X0The j characteristic item that () is current goal event time series on jth attribute, | X0(j) | for its length, XkJ () is kth case time series on jth attribute, | Xk(j) | for its length.
Manhatton distance belongs to distance calculating method two kinds different with consolidation path distance, and generally, consolidation path distance may more than manhatton distance, but they have been comprised in comprehensive distance Sim simultaneouslykIn.For above-mentioned situation, it is possible to by the adjustment to case characteristic item weight, both distances are made can the match condition of each characteristic item to be reflected in comprehensive distance better when carrying out cumulative.
Step 3: case is reused, method particularly includes:
Case is reused and is comprised determining whether that the match is successful and case result two steps of output.If manufacturing industry scene current goal event and the comprehensive distance of certain case in case library are less than setpoint distance threshold value, then the match is successful with this case for current goal event, both have close feature, and it may happen that identical result, the case result of this case is exported and is used for object event, think that current manufacturing industry scene there occurs the ruuning situation that case result is consistent, then perform next step；Otherwise, it fails to match and terminates；
Step 4: Case-based adaptation, method particularly includes:
Case-based adaptation includes the concordance of case result and manufacturing industry scene current goal event and judges and two steps of case modified result.The concordance judgement of case result and current goal event refers to after case result is exported object event, judge whether object event there occurs the case result of correspondence really, if object event there occurs the case result of correspondence really, then case result is consistent with current goal event；Otherwise case result is inconsistent with current goal event.Such as " motor damage " this case result is applied in " motor status judgement " this object event, namely think that current motor has damaged and taked the solution of correspondence, if find out that motor damages really afterwards, then case result is consistent with current goal event；Otherwise case result is inconsistent with current goal event.
If case result is consistent with current goal event, then this case is correct, and reasoning by cases terminates；Otherwise this case mistake, is modified the case result that the match is successful according to the feedback information of object event, and revised case result is submitted to case library, and reasoning by cases terminates.
Step 5: case characteristic item weight adjusts, method particularly includes:
The initial weight of case characteristic item, the i.e. initial weight of each monitoring parameter that case is corresponding, by manufacture field, expert directly gives, and has certain subjective experience, so needing constantly Feature item weighting to be adjusted in daily process operation so that it is more have engineering practicability.
The adjustment flow process of the case characteristic item weight of present invention design is as shown in Figure 2, the present invention adjusts in amplitude formula in weight and adds matching distance proportion coefficients R, proportion coefficients R can adopt different expression formulas according to the size of the proportion that characteristic item matching distance accounts for comprehensive distance, mainly serves for ensuring following 2 points:
(1) when characteristic item matching distance accounts for the proportion of comprehensive distance more than or equal to characteristic item threshold value TDistTime, characteristic item is more likely to reduce the similarity of case in current goal event and case library, then weight adjusts amplitude with characteristic item matching distance is proportional relation；
(2) when characteristic item matching distance accounts for the proportion of comprehensive distance less than characteristic item threshold value TDistTime, characteristic item is more likely to increase the similarity of case in current goal event and case library, then weight adjusts amplitude with characteristic item matching distance is inverse relation.
Setting of matching distance proportion coefficients R on the one hand can so that the correction of weight more conforms to the cognitive law of the mankind, and the addition of matching distance proportion coefficients R also solves the numerical value variability issues of dynamic time warping distance and manhatton distance on the other hand.
When adjusting Feature item weighting, being incorporated in formula by matching distance proportion coefficients R, it is as follows that each weight adjusts amplitude formula:
W'kj=Wkj(1±Δ×Rkj)(3)
W in above formulakjRepresent characteristic item j weight before correction in case k, W'kjRepresenting that in case k, characteristic item j is in revised weight, Δ represents the range coefficient of adjustment every time, RkjRepresenting the matching distance proportion coefficients of characteristic item j in case k, computing formula is as follows:
In formula (4), SimkRepresent the comprehensive distance of kth case and current goal event, T in case libraryDistRepresent characteristic item threshold value, Dist (X0(j),Xk(j)) represent kth case and current goal sequence distance on jth attribute, according to R in formula (4)kjComputing formula, it can be deduced that formula (3) altogether has following four kinds of situations:
(1) when case k is correct and Dist (X0(j),Xk(j)/Simk>=TDistTime, reduce the weight of case item j, W' in case kkj=Wkj(1-Δ×ωkjDist(X0(j),Xk(j)/Simk), and Dist (X0(j),XkJ () is more big, reduce amplitude more big；
(2) when case k is correct and Dist (X0(j),Xk(j)/Simk< TDistTime, increase the weight of case item j, W' in case kkj=Wkj(1+Δ×(1-ωkjDist(X0(j),Xk(j)/Simk)), and Dist (X0(j),XkJ () is more little, the amplitude of increase is more big；
(3) when case k is judged as mistake and Dist (X0(j),Xk(j)/Simk>=TDistTime, increase the weight of case item j, W' in case kkj=Wkj(1+Δ×ωkjDist(X0(j),Xk(j)/Simk), and Dist (X0(j),XkJ () is more big, the amplitude of increase is more big；
(4) when case k is judged as mistake and Dist (X0(j),Xk(j)/Simk< TDistTime, reduce the weight of case item j, W' in case kkj=Wkj(1-Δ×(1-ωkjDist(X0(j),Xk(j)/Simk)), and Dist (X0(j),XkJ () is more little, the amplitude of reduction is more big.
The flow process of the reasoning by cases method of the support Time Series Matching that Fig. 1 provides for the embodiment of the present invention, concrete method for designing includes following five steps:
1) case structure
For electromotor, setting up following three groups of cases, often group case all includes case title, case result, case effect assessment and case characteristic item, as shown in table 1 to table 3.
Choosing voltage, electric current and temperature and be used as case characteristic item, wherein temperature profile item supports time point Value Types and time series type simultaneously, and voltage and current only supports time point Value Types.
Case effect assessment is based on case afterwards and reuses effect and carry out adding typing, so first being left a blank during structure case.
Table 1 electromotor case 1
Table 2 electromotor case 2
Table 3 electromotor case 3
2) Case Retrieval
The voltage of input current time is 190, electric current is 32, temperature is 180, wherein temperature historical trend before current time is 50,55,80,110,140,155,163,170, have benefited from dynamic time warping distance algorithm, it is possible to support that the distance of the unequal length between current time sequence and history case characteristic item time series is mated.Case Retrieval result is as shown in table 4.
Table 4 Case Retrieval result
| Case title | Similarity mode distance |
| Electromotor case 1 | 32.6 |
| Electromotor case 2 | 59 |
| Electromotor case 3 | 110.4 |
In table 4, the distance that electromotor case 1 and electromotor case 3 all calculate is manhatton distance distance, and the distance that electromotor case 2 calculates is dynamic time warping distance.
Sort from high to low according to the similarity with current goal case, be followed successively by electromotor case 1, electromotor case 2 and electromotor case 3.The numerical value that voltage in electromotor case 1, electric current are corresponding with current goal case with temperature these three characteristic item numerical value is very close, and therefore similarity mode is apart from only small, and similarity is significantly high；Although the temperature-time sequence in electromotor case 2 is inconsistent with the seasonal effect in time series length of current goal case, but trend is close, is the process of sharp increase, and therefore similarity mode distance is less, and similarity is higher；The numerical value difference that various features item numerical value in electromotor case 3 is corresponding with current goal case is relatively big, and therefore similarity mode distance is relatively big, and similarity is relatively low.
3) case is reused
Set the similarity mode distance threshold of reasoning by cases as 80, then based on the Case Retrieval result of step 2, it is considered for electromotor case 1 and electromotor case 2 carries out case and reuses, think due to overtension, electric current is too low, the reasons such as constant temperature increases, electromotor burns out at present, it is necessary to electromotor is looked into and repaiies.
4) Case-based adaptation
Finding through expert's field verification, electromotor burns out really, and namely corresponding case result is consistent with object event, is therefore updated to excellent by the case effect assessment of electromotor case 1 and electromotor case 2, and without the case reused is modified.
5) case characteristic item weight adjusts
Through expert judgments, it is best that electromotor case 1 reuses effect, carries out weight adjustment hence for case 1, Δ is set to 0.1, TDistIt is set to 0.5.In electromotor case 1, it is as follows that the weight of voltage, electric current and temperature these three characteristic item adjusts process.
(1) proportion that the matching distance of this characteristic item of voltage is shared in comprehensive distance is relatively big, is 0.9 × (220-190)/32.6=0.828, more than TDist, it is therefore desirable to reducing its Feature item weighting, revised weight is 0.9 × (1-0.1 × 27/32.6)=0.825.
(2) proportion that the matching distance of this characteristic item of electric current is shared in comprehensive distance is 0.7 × (40-32)/32.6=0.172, less than TDist, so improving its Feature item weighting, revised weight is 0.7 × (1+0.1 × (1-0.172))=0.758.
(3) matching distance of this characteristic item of temperature is 0, and proportion shared in comprehensive distance is less than TDist, so improving its Feature item weighting, revised weight is 0.8 × (1+0.1 × 1)=0.88.
After above-mentioned weight correction, it is possible to make the weight of case characteristic item more conform to practical situation, improve the accuracy of reasoning by cases in the future.
Claims (10)
1. the reasoning by cases method supporting Time Series Matching, it is characterised in that comprise the following steps:
1) structure case library；
2) Case Retrieval: with the status information of current goal event for input, Design case based storehouse carries out Case Retrieval, obtains current goal event and the comprehensive distance of each case in case library；
3) case is reused: if current goal event is with case library, the comprehensive distance of certain case is less than setpoint distance threshold value, then the match is successful with this case for current goal event, the case result of this case is exported and is used for object event, performs next step；Otherwise, it fails to match and terminates；
4) Case-based adaptation: judge that whether the case result that the match is successful is consistent with current goal event；If consistent, then this case is correct, and reasoning by cases terminates；Otherwise this case mistake, is modified the case result that the match is successful according to the feedback information of object event, and revised case result is submitted to case library, and reasoning by cases terminates.
2. a kind of reasoning by cases method supporting Time Series Matching according to claim 1, it is characterised in that described case library is the set of multiple case, and described case includes: case title, case characteristic item set, case result, case effect assessment.
3. a kind of reasoning by cases method supporting Time Series Matching according to claim 1, it is characterised in that described Case Retrieval formula is as follows:
In above formula, SimkRepresent kth case X in case librarykCharacteristic item sequence X with current goal event0Comprehensive distance, ωkjFor the weight that the jth attribute of kth case is shared in participating in the ATTRIBUTE INDEX of case coupling, j=1,2 ..., m, m is the attribute number of kth case；Dist(X0(j),Xk(j)) represent the characteristic item sequence X of kth case and current goal event0Matching distance on jth attribute.
4. a kind of reasoning by cases method supporting Time Series Matching according to claim 3, it is characterised in that the characteristic item of described object event is time point value, then Dist (X on jth attribute0(j),Xk(j)) for manhatton distance | X0(j),Xk(j)|；X0The j characteristic item sequence that () is current goal event time point value on jth attribute, XkJ () is kth case time point value on jth attribute.
5. a kind of reasoning by cases method supporting Time Series Matching according to claim 3, it is characterised in that the characteristic item of described object event is time series on jth attribute, Dist (X0(j),Xk(j)) in dynamic time warping distance algorithm consolidation path distance D (| Xk(j)|,|X0(j)|)；X0The j characteristic item sequence that () is current goal event time series on jth attribute, XkJ () is kth case time series on jth attribute.
6. a kind of reasoning by cases method supporting Time Series Matching according to claim 1, it is characterised in that also include the weight of case characteristic item is adjusted, adjusts amplitude formula by weight and realizes:
W′kj=Wkj(1±Δ×Rkj)(2)
W in above formulakjRepresent characteristic item j weight before correction in case k, W 'kjRepresenting that in case k, characteristic item j is in revised weight, Δ represents the range coefficient of adjustment every time, RkjRepresenting the matching distance proportion coefficients of characteristic item j in case k, computing formula is as follows:
In above formula, SimkRepresent kth case X in case librarykCharacteristic item sequence X with current goal event0Comprehensive distance, TDistRepresent characteristic item matching distance threshold value, Dist (X0(j),Xk(j)) represent the characteristic item sequence X of kth case and current goal event0Matching distance on jth attribute；ωkjFor the weight that the jth attribute of kth case is shared in the ATTRIBUTE INDEX participating in case coupling.
7. a kind of reasoning by cases method supporting Time Series Matching according to claim 6, it is characterised in that: when case k is correct and Dist (X0(j),Xk(j)/Simk>=TDistTime:
W′kj=Wkj(1-Δ×ωkjDist(X0(j),Xk(j)/Simk)。
8. a kind of reasoning by cases method supporting Time Series Matching according to claim 6, it is characterised in that when case k is correct and Dist (X0(j),Xk(j)/Simk< TDistTime:
W′kj=Wkj(1+Δ×(1-ωkjDist(X0(j),Xk(j)/Simk))。
9. a kind of reasoning by cases method supporting Time Series Matching according to claim 6, it is characterised in that when case k is judged as mistake and Dist (X0(j),Xk(j)/Simk>=TDistTime:
W′kj=Wkj(1+Δ×ωkjDist(X0(j),Xk(j)/Simk)。
10. a kind of reasoning by cases method supporting Time Series Matching according to claim 6, it is characterised in that when case k is judged as mistake and Dist (X0(j),Xk(j)/Simk< TDistTime:
W′kj=Wkj(1-Δ×(1-ωkjDist(X0(j),Xk(j)/Simk))。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410797922.XA CN105787610A (en) | 2014-12-18 | 2014-12-18 | Case-based reasoning method capable of supporting time sequence matching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410797922.XA CN105787610A (en) | 2014-12-18 | 2014-12-18 | Case-based reasoning method capable of supporting time sequence matching |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105787610A true CN105787610A (en) | 2016-07-20 |
Family
ID=56384993
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410797922.XA Pending CN105787610A (en) | 2014-12-18 | 2014-12-18 | Case-based reasoning method capable of supporting time sequence matching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105787610A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599264A (en) * | 2016-12-22 | 2017-04-26 | 国家行政学院 | Big data-based emergency event evolution reasoning method and system |
| CN106897378A (en) * | 2017-01-20 | 2017-06-27 | 浙江大学 | A kind of semantic parameter search dynamic regulating method towards three-dimensional CAD model |
| CN107358348A (en) * | 2017-07-05 | 2017-11-17 | 成都电科智联科技有限公司 | A kind of method that plan fault-tolerance is improved based on OEE |
| CN108090567A (en) * | 2018-01-19 | 2018-05-29 | 国家电网公司 | Power communication system method for diagnosing faults and device |
| CN108231137A (en) * | 2016-12-15 | 2018-06-29 | 童综合医疗社团法人童综合医院 | A kind of medical system with feedback study |
| CN108305007A (en) * | 2018-02-05 | 2018-07-20 | 深圳市瑞捷建筑工程咨询有限公司 | Engineering construction report-generating method, device and its system, server |
| CN108764805A (en) * | 2018-06-11 | 2018-11-06 | 河南理工大学 | A kind of multi-model self-adapting recommendation method and system of collaborative logistics Services Composition |
| CN108830306A (en) * | 2018-05-30 | 2018-11-16 | 东软集团股份有限公司 | The workflow method for diagnosing faults and device, medium and electronic equipment of business datum |
| CN109556861A (en) * | 2019-01-23 | 2019-04-02 | 昆明理工大学 | A kind of bearing real-time fault diagnosis system of case-based reasioning |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101383027A (en) * | 2008-09-12 | 2009-03-11 | 同济大学 | Environmental emergency scheme generating method and system |
| CN102176239A (en) * | 2011-02-23 | 2011-09-07 | 电子科技大学 | Kitchenware fault diagnosing method based on case-based reasoning |
| CN103135014A (en) * | 2012-12-14 | 2013-06-05 | 西安电子科技大学 | Transformer fault diagnosis method based on case-based reasoning |
-
2014
- 2014-12-18 CN CN201410797922.XA patent/CN105787610A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101383027A (en) * | 2008-09-12 | 2009-03-11 | 同济大学 | Environmental emergency scheme generating method and system |
| CN102176239A (en) * | 2011-02-23 | 2011-09-07 | 电子科技大学 | Kitchenware fault diagnosing method based on case-based reasoning |
| CN103135014A (en) * | 2012-12-14 | 2013-06-05 | 西安电子科技大学 | Transformer fault diagnosis method based on case-based reasoning |
Non-Patent Citations (3)
| Title |
|---|
| 周荣良: "装备车辆故障的远程诊断", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 徐小力，洪孝安主编: "《设备与维修工程现代技术及应用》", 31 January 2009 * |
| 程文聪: "面向大规模网络安全态势分析的时序数据挖掘关键技术研究", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108231137A (en) * | 2016-12-15 | 2018-06-29 | 童综合医疗社团法人童综合医院 | A kind of medical system with feedback study |
| CN106599264A (en) * | 2016-12-22 | 2017-04-26 | 国家行政学院 | Big data-based emergency event evolution reasoning method and system |
| CN106599264B (en) * | 2016-12-22 | 2019-07-19 | 国家行政学院 | A kind of emergency event evolution inference method and system based on big data |
| CN106897378A (en) * | 2017-01-20 | 2017-06-27 | 浙江大学 | A kind of semantic parameter search dynamic regulating method towards three-dimensional CAD model |
| CN107358348A (en) * | 2017-07-05 | 2017-11-17 | 成都电科智联科技有限公司 | A kind of method that plan fault-tolerance is improved based on OEE |
| CN108090567A (en) * | 2018-01-19 | 2018-05-29 | 国家电网公司 | Power communication system method for diagnosing faults and device |
| CN108305007A (en) * | 2018-02-05 | 2018-07-20 | 深圳市瑞捷建筑工程咨询有限公司 | Engineering construction report-generating method, device and its system, server |
| CN108830306A (en) * | 2018-05-30 | 2018-11-16 | 东软集团股份有限公司 | The workflow method for diagnosing faults and device, medium and electronic equipment of business datum |
| CN108764805A (en) * | 2018-06-11 | 2018-11-06 | 河南理工大学 | A kind of multi-model self-adapting recommendation method and system of collaborative logistics Services Composition |
| CN109556861A (en) * | 2019-01-23 | 2019-04-02 | 昆明理工大学 | A kind of bearing real-time fault diagnosis system of case-based reasioning |
| CN109556861B (en) * | 2019-01-23 | 2020-08-07 | 昆明理工大学 | Bearing real-time fault diagnosis system based on case reasoning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105787610A (en) | Case-based reasoning method capable of supporting time sequence matching | |
| US11093314B2 (en) | Time-sequential data diagnosis device, additional learning method, and recording medium | |
| US20050273296A1 (en) | Neural network model for electric submersible pump system | |
| CN105787563A (en) | Self-learning mechanism-base fast matching fuzzy reasoning method | |
| CN113887846B (en) | Out-of-tolerance risk early warning method for capacitive voltage transformer | |
| Duer | Applications of an artificial intelligence for servicing of a technical object | |
| CN112098848A (en) | Battery discharge capacity prediction method and system and readable storage medium | |
| CN113159341A (en) | Power distribution network aid decision-making method and system integrating deep reinforcement learning and expert experience | |
| CN104715342A (en) | Electrical equipment fault handling and aid decision making method based on case-based reasoning | |
| CN113837596A (en) | Fault determination method and device, electronic equipment and storage medium | |
| CN114626769B (en) | Operation and maintenance method and system for capacitor voltage transformer | |
| CN104062541A (en) | Fault diagnosis method for stator turn-to-turn short circuit of fault-tolerant permanent magnet motor | |
| CN111192158A (en) | Transformer substation daily load curve similarity matching method based on deep learning | |
| CN111680712A (en) | Transformer oil temperature prediction method, device and system based on similar moments in the day | |
| CN114065114A (en) | Method and system for predicting metering error of capacitive voltage transformer | |
| CN109961160B (en) | Power grid future operation trend estimation method and system based on tide parameters | |
| Liu et al. | A multivariate forecasting method for short‐term load using chaotic features and RBF neural network | |
| Khalyasmaa et al. | Training samples construction for energy utilities operational assets management | |
| CN116736214A (en) | CVT stability evaluation method and system | |
| CN115759455A (en) | Load probability density prediction method based on time sequence Gaussian mixture density network | |
| CN112906130A (en) | Structure reliability assessment method based on small sample data | |
| CN114977953A (en) | Motor output torque correction method, device and medium based on neural network | |
| CN112598259A (en) | Capacity measuring method and device and computer readable storage medium | |
| Al-Romaimi et al. | Strategic Asset Management Health Index for Predicting Power Transformer Health Conditions | |
| Maragatham et al. | Predicting the Uncertainty of Data Generated by Power Transformer using Gradient Boosting Decision Tree |
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: 20160720 |