CN110599122A - Power grid dispatching system page recommendation method based on pattern mining and correlation analysis - Google Patents
Power grid dispatching system page recommendation method based on pattern mining and correlation analysis Download PDFInfo
- Publication number
- CN110599122A CN110599122A CN201910812889.6A CN201910812889A CN110599122A CN 110599122 A CN110599122 A CN 110599122A CN 201910812889 A CN201910812889 A CN 201910812889A CN 110599122 A CN110599122 A CN 110599122A
- Authority
- CN
- China
- Prior art keywords
- page
- session
- transaction
- frequent
- transactions
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a power grid dispatching system page recommendation method based on pattern mining and association analysis.
Description
Technical Field
The invention relates to the technical field of automation of a power grid dispatching system, in particular to a power grid dispatching system page recommendation method based on pattern mining and correlation analysis.
Background
With the expansion of the scale of the power grid and the improvement of the power supply quality requirements of users, the power grid dispatching operation is increasingly complex, fine and standard.
For a common scheduling service, a dispatcher generally needs to browse and analyze multiple indexes of multiple system pages, and make a reasonable treatment measure by combining actual conditions of each device.
Usually, a task is completed, and a dispatcher needs to perform complicated page browsing and data analysis, so that the workload is large and the efficiency is low.
Disclosure of Invention
The invention aims to provide a power grid dispatching system page recommending method based on mode mining and correlation analysis, which is used for carrying out page navigation or intelligent recommendation in the daily browsing process of a dispatcher on the basis of mining an operation mode of the dispatcher in daily work.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the power grid dispatching system page recommendation method based on pattern mining and correlation analysis comprises the following steps:
acquiring an operation record of a dispatcher, and dividing a session; the operation records of the dispatcher are a series of page sequences which are continuous in time;
performing transaction identification on the divided sessions to obtain a transaction sequence of each session;
acquiring a frequent mode matched with the transaction sequence;
and acquiring the rest transactions except the matching items in the frequent pattern with the pattern matching relationship with the real-time transaction sequence as recommended transactions.
Further, the performing session division includes:
the conversation division is carried out by adopting a time window method, if the time interval between two continuous pages is greater than a certain threshold value, the original conversation is cut off, and a new conversation is generated; the time interval between two consecutive pages is defined as the dwell time of the previous page.
Further, the transaction identification of the divided session by using a reference length method includes:
reading page browsing records in sequence, and determining as a transaction once a navigation page is converted into a content page;
the navigation page and the content page are defined as follows:
when the transaction reference value of a certain page exceeds a set threshold value, the page is a content page; otherwise, the page is a navigation page;
the transaction reference value refers to the upper limit of the browsing duration that 70% of the page instances do not exceed in the page instances with the dwell time less than 600 s.
Further, the method also comprises a step of performing hierarchical division on the page subjected to the transaction identification, and the method comprises the following steps:
different system diagrams are defined by adopting different page numbers; for the station graph and the curve graph, pages with the same page type but different specific examples are normalized, and the same page number is defined according to the page type.
Further, acquiring a frequent pattern matching the transaction sequence includes:
adopting a neighbor propagation clustering algorithm to carry out frequent pattern mining;
adopting an FP _ Growth algorithm to mine frequent items in a frequent mode; the frequent item is a transaction contained in the session in the frequent mode;
if there is an intersection between the obtained transaction sequence and the frequent item in the frequent pattern, the frequent pattern is a matching pattern of the current real-time transaction sequence.
Further, a neighbor propagation clustering algorithm is adopted to perform frequent pattern mining, which comprises the following steps:
establishing a session vector for the two sessions according to a 0-0 matching principle, calculating session similarity, and establishing a session similarity matrix;
clustering the sessions by adopting a neighbor propagation clustering algorithm, wherein each class corresponds to a frequent mode;
and determining the optimal clustering result by taking the contour coefficient as a standard for evaluating the clustering effect and adjusting the attenuation coefficient.
Further, establishing a session vector for the two sessions according to a 0-0 matching principle, including:
taking a transaction union set of the session 1 and the session 2 to form a transaction vector;
representing elements in the session vector by the number of times a transaction in the transaction vector occurs in the session;
calculating the conversation similarity as follows:
where s (x, y) represents the similarity of session x and session y.
Further, in the clustering process, an attenuation coefficient lambda is introduced:
rt+1(i,k)←(1-λ)rt+1(i,k)+λrt(i,k)
at+1(i,k)←(1-λ)at+1(i,k)+λat(i,k)
where r (i, k) represents an element in the attraction matrix, a (i, k) is an element in the attribution matrix, i represents the ith session, k represents the kth session, and subscript t represents the tth iteration.
Further, acquiring the remaining transactions except the matching items in the frequent pattern having the pattern matching relationship with the real-time transaction sequence as recommended transactions, further comprising:
calculating confidence degrees of the recommended transactions one by one, and selecting the transactions N before the confidence degree as recommended contents;
the confidence coefficient is as follows: probability of simultaneous occurrence of Y transactions on the premise of occurrence of X transactions; the calculation method is the ratio between the number of transactions containing X and Y transactions and the number of transactions containing X.
Further, when the recommended transaction is an independent page, the independent page is given as a recommendation result; when the recommended transaction is a page class, a specific instance page of the page class needs to be given as a recommendation result.
The invention achieves the following beneficial effects:
the method and the system can dig out the operation mode of power grid dispatching and intelligently recommend the page needing to be browsed by the dispatcher to a specific mode, so that the working efficiency of the dispatcher can be obviously improved. The method is used as a typical application for supporting the intellectualization of the scheduling operation, and has important significance for improving the intellectualization level of the scheduling system.
Drawings
FIG. 1 is a flowchart of a method for recommending a power grid scheduling system page based on pattern mining and correlation analysis according to the present invention;
FIG. 2 is a diagram of transactions in an embodiment of the invention;
FIG. 3 is a diagram of a FP tree with head pointers according to an embodiment of the present invention.
Detailed Description
The invention is further described below. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The embodiment of the invention provides a power grid dispatching system page recommendation method based on pattern mining and association analysis, which is realized through a data preprocessing module, a cluster mining module, an association analysis module and an intelligent recommendation module.
A data preprocessing module:
11. session partitioning
The dispatcher's operating record is a temporally continuous series of page sequences, in which different operating modes are included. The conversation division divides the whole series of page sequences into blocks according to different operation purposes, and each page sequence is a conversation for a dispatcher to complete a certain task.
The conversation division is carried out by adopting a time window method. The time window method considers that when the time interval between two continuous pages is larger than a certain threshold value, the original conversation is cut off, and a new conversation is generated. Wherein, the time interval of two continuous pages is defined as the stay time of the previous page.
Typically the session split threshold defaults to 600 s. In the power grid dispatching system, in order to monitor certain indexes, regulating personnel often stay on some pages for a long time, which is far more than 600s, and the conversation is not disconnected. Aiming at the characteristics, the session division threshold values of different types of pages need to be dynamically adjusted.
The invention gives different conversation division thresholds by counting the distribution intervals of most of the example stay time in each page. The threshold that the dwell time of 70% of the page instances does not exceed is defined as the page reference time. When the reference time length of the page exceeds a certain range, the dispatcher is considered to stay in the page frequently for monitoring, the conversation division threshold value of the page of the type is larger than the default value, and the conversation division threshold value is recommended to be adjusted to be twice of the default value. The proportionality coefficient of 70% can be adjusted according to the data situation.
12. Transaction identification
Several steps are included in each session that are necessary to complete the task.
A transaction refers to a sequence of pages within a session that are browsed for completion of a single step in the task, with one or more transactions in a session.
The invention adopts a reference length method to identify the affairs. The reference length method divides a page into a content page and a navigation page. The content page is a page which has an access purpose and is interested by a regulatory person, and the navigation page is a page with the jumping property of the regulatory person in the browsing process. The page browsing records are sequentially read, and once the navigation page is converted into the content page, the boundary of the transaction is determined. The transaction is shown in FIG. 2, where N represents a navigation page and C represents a content page.
The transaction identification is performed by first analyzing the transaction reference value. And defining the transaction reference value as an upper limit of the browsing duration which is not exceeded by 70% of page instances with the dwell time less than 600 s. When the transaction reference value of a certain page exceeds a set threshold value, the page is a content page; otherwise the page is a navigation page. The set threshold is given by comparing the transaction reference values of the typical content page and navigation page.
13. Page hierarchy partitioning
In the power grid dispatching system, some types of pages can show the contents of different stations and equipment. For example, all "XX substations. fac" represent specific plant diagrams, which are plant-type pages. The page needs to be normalized in the operation mode mining, so that the influence on the subsequent mode clustering effect caused by the fact that specific station and equipment information are mixed in the operation mode is avoided.
The types of the pages in the scheduling system can be generally divided into types such as a system diagram, an application diagram, a plant station diagram, a curve diagram and the like. Therefore, for a system diagram with uniqueness logically, page hierarchy division defines different system diagrams by adopting different page numbers; and the pages with the same page type but different specific examples, such as a station diagram, a curve diagram and the like, are subjected to normalization processing, and are defined as the same page number according to the page type.
A clustering mining module:
and (4) carrying out pattern mining by adopting an Adjacent Propagation (AP) clustering algorithm. The algorithm does not need to appoint cluster number of clusters in advance, the input parameters are the similarity between every two sessions, and the problem that the vector dimensions of a plurality of sessions cannot be unified can be solved.
21. Session similarity calculation
The input parameter of the AP cluster is a similarity matrix, and the similarity between sessions is calculated as follows:
first, a session vector is established for two sessions according to the 0-0 matching principle. Assume that session 1 is (a, b, c, b) and session 2 is (b, c, d, c, e, f). The union of the transactions of session 1 and session 2 is (a, b, c, d, e, f). Session 1 and session 2 may be represented by vectors (x1, x2, x3, x4, x5, x6) according to the union of transactions. Where x1 represents the number of times a transaction occurs in a session, and so on. The vector representations of final session 1 and session 2 are (1, 2, 1, 0, 0, 0) and (0, 1, 2, 1, 1, 1), respectively.
The session similarity is then calculated. If x and y are two session vectors, the session similarity is:
22. AP clustering
AP clustering is a clustering algorithm based on information transfer between data points. The AP algorithm is carried out by iteratively updating two matrixes of the attraction degree and the attribution degree:
r (i, k) describes the degree to which a data object is suitable as a cluster center, and the elements in the attraction matrix iterate according to the following formula:
where a (i, k) is an element in the attribution degree matrix, and represents the degree to which the ith session object selects the kth session object as its representative, and s (i, k) is the similarity between the ith session object and the kth session object. The initialized value of the element in the attribution degree matrix is 0.
A (i, k) describes the suitability of the data object to select other objects as its data clustering center, and the elements in the attribution matrix iterate according to the following formula:
in order to avoid oscillation, the AP algorithm introduces a damping coefficient lambda when updating information,
rt+1(i,k)←(1-λ)rt+1(i,k)+λrt(i,k)
at+1(i,k)←(1-λ)at+1(i,k)+λat(i,k)
if the algorithm remains unchanged after a number of iterations or if the algorithm execution exceeds a set number of iterations, the algorithm ends.
23. Clustering result evaluation
The method evaluates the clustering result through the contour coefficient. The contour coefficients for the individual points are:
si＝(bi-ai)/max(ai,bi)
wherein, biIs the minimum of the average distance of the ith object and all objects in any cluster not containing the object, aiThe average distance from the ith object to all other objects in the cluster to which it belongs.
The contour coefficients for all points are represented by the average of the individual point contour coefficients.
Association analysis module
The invention adopts FP _ Growth algorithm to discover a frequent item set. The FP _ Growth algorithm comprises two steps of building a FP tree and mining a frequent item set.
31. FP tree construction
In constructing the FP-tree, the dataset needs to be scanned twice, the first scan to count the frequency, the second scan to construct a frequent item tree, only considering frequent transactions.
The FP-tree is illustrated below by way of example. Assume that there is one cluster in which the sample of session data is as follows:
TABLE 1 sessions in a cluster
| Session ID | Transactions in a session |
| 1 | r,z,h,j,p |
| 2 | z,y,x,w,v,u,t,s |
| 3 | z |
| 4 | r,x,n,o,s |
| 5 | y,r,x,z,q,t,p |
| 6 | y,z,x,e,q,s,t,m |
The definition support degree is as follows:
Support(S)＝m/n
wherein m is the support count of the transaction S, i.e. the number of occurrences of the transaction S in all session sequences; n is the number of all sessions;
first, a head pointer table is constructed, and all elements of a given type in the FP-tree can be quickly accessed by using the head pointer table.
The head pointer table is shown in table 2 below.
Watch 2 head pointer
The affairs in the head pointer table are affairs which are larger than the minimum support degree in the conversation sample, if the support degree of the affairs S is larger than a given minimum support degree threshold value, the affairs S are frequent affairs in the conversation set; the transaction count is the frequency of occurrence of transactions in the session sample.
In addition, the head pointer table also stores the information that the transaction node points to the corresponding instance on the FP tree, and the information is used for quickly accessing all the elements in the FP tree.
And then starting to build the FP tree, and firstly removing the transactions which do not meet the minimum support degree in the session. The transactions in each session are then ordered based on their absolute frequency of occurrence.
The sessions in table 1 are filtered and reordered as shown in table 3 below.
TABLE 3 reordered sessions
| Session ID | Transactions in a session | Filtered and reordered transactions |
| 1 | r,z,h,j,p | z，r |
| 2 | z,y,x,w,v,u,t,s | z，x，y，s，t |
| 3 | z | z |
| 4 | r,x,n,o,s | x，s，r |
| 5 | y,r,x,z,q,t,p | z，x，y，r，t |
| 6 | y,z,x,e,q,s,t,m | z，x，y，s，t |
And constructing the FP tree in the next step. Taking the empty set as a root node, and sequentially adding the transaction items subjected to traversal filtering and sequencing into a tree, wherein if existing elements exist in a fruit tree, the count value of the existing elements is increased; if an existing element does not exist, a branch is added to the tree. The final FP-tree with head pointers is constructed as in fig. 3.
32. Mining frequent item sets
After the FP tree is constructed, a frequent item set can be extracted, and the method comprises the following three steps:
1) obtaining a conditional pattern base from the FP tree;
the conditional mode base is a set of paths ending with the element entry sought, representing all the content between the element entry sought and the tree root node. For example, in FIG. 3, the prefix paths for element r are x, s, z, x, y, and z. Each prefix path is associated with a count value that indicates the number of r paths on each path. In order to obtain these prefix paths, all the type element nodes on the FP-tree can be sequentially obtained according to the start pointer containing the type element in the head pointer table through the previously obtained head pointer table. The tree is then traced up to the root node according to each element item.
2) After obtaining the conditional mode base of each element (transaction in the embodiment of the present invention), it is determined whether the path set is a frequent item according to the count value. In addition, a conditional FP-tree is created for each element and its conditional schema base.
3) Iteratively repeating steps 1) and 2), recursively finding frequent items, finding conditional pattern bases, and finding additional conditional trees until a tree contains an element item.
The intelligent recommendation module:
41. transaction recommendation
After the frequent pattern mining is completed, the transaction recommendation can be performed based on the frequent pattern and the FP tree, and the method comprises the following specific steps:
1) and carrying out session division and transaction identification in real time, and obtaining the current ongoing session of the dispatcher and a corresponding transaction sequence.
2) And traversing the frequent patterns of each behavior class, and performing pattern matching. And if the real-time transaction sequence and the frequent pattern of the decision behavior class have intersection, the frequent pattern is a matching pattern of the current real-time transaction sequence. The behavior classes are user behavior patterns mined in cluster mining, and generally one cluster corresponds to one behavior class.
Frequent association transaction items in each clustering pattern can be obtained by mining frequent item sets, and a group of frequent transaction items with association corresponds to a frequent pattern.
3) And identifying a frequent pattern with a pattern matching relation with the real-time transaction sequence, and selecting the rest transactions except the matching items in the frequent pattern as alternative recommended transactions. And calculating the confidence degrees of the alternative recommended transactions one by one, and selecting the transaction items N times before the confidence degree as recommended contents. For example: if the real-time sequence is (a, e, b) and the frequent pattern is (a, b, d), the real-time sequence and the frequent pattern have matching items (a, b), and the remaining transactions d in the frequent pattern can be used as recommended items.
Defining confidence: the probability that Y transactions occur simultaneously is given by the premise that X transactions occur. The calculation method is the ratio between the number of transactions containing X and Y and the number of transactions containing X.
42. Type page instantiation
Generally, there are two types of transactions recommended, one is a separate page and one is a page class. For example, the substation summary view is an independent page showing all substation entry links; the substation view is a page class, and the page can show different substations.
When the recommended transaction is a separate page, the separate page may be given as a result of the recommendation. When the recommended transaction is a page class, a specific instance page of the page class needs to be given as a recommendation result. For example, through association mining, if the recommendation output to the user is a substation view, it needs to be further determined which substation view corresponds to specifically.
The invention carries out instantiation page recommendation by using a statistical method. For example, the recommended transaction is a graph view. And counting the frequency ranking of the curve views in the clustering mode, and selecting the instantiation curve view with the highest frequency of occurrence as an instantiation recommendation result.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (10)
1. The power grid dispatching system page recommendation method based on pattern mining and correlation analysis is characterized by comprising the following steps of:
acquiring an operation record of a dispatcher, and dividing a session; the operation records of the dispatcher are a series of page sequences which are continuous in time;
performing transaction identification on the divided sessions to obtain a transaction sequence of each session;
acquiring a frequent mode matched with the transaction sequence;
and acquiring the rest transactions except the matching items in the frequent pattern with the pattern matching relationship with the real-time transaction sequence as recommended transactions.
2. The grid scheduling system page recommendation method based on pattern mining and correlation analysis of claim 1, wherein the performing session division comprises:
the conversation division is carried out by adopting a time window method, if the time interval between two continuous pages is greater than a certain threshold value, the original conversation is cut off, and a new conversation is generated; the time interval between two consecutive pages is defined as the dwell time of the previous page.
3. The method for recommending a page of a power grid scheduling system based on pattern mining and correlation analysis according to claim 1, wherein the step of performing transaction recognition on the divided conversation by using a reference length method comprises the following steps:
reading page browsing records in sequence, and determining as a transaction once a navigation page is converted into a content page;
the navigation page and the content page are defined as follows:
when the transaction reference value of a certain page exceeds a set threshold value, the page is a content page; otherwise, the page is a navigation page;
the transaction reference value refers to the upper limit of the browsing duration that 70% of the page instances do not exceed in the page instances with the dwell time less than 600 s.
4. The method for recommending a power grid scheduling system page based on pattern mining and correlation analysis according to claim 1, further comprising a step of performing hierarchical division on the page subjected to transaction recognition, including:
different system diagrams are defined by adopting different page numbers; for the station graph and the curve graph, pages with the same page type but different specific examples are normalized, and the same page number is defined according to the page type.
5. The method for recommending a page of a power grid scheduling system based on pattern mining and correlation analysis according to claim 1, wherein obtaining frequent patterns matching the transaction sequence comprises:
adopting a neighbor propagation clustering algorithm to carry out frequent pattern mining;
adopting an FP _ Growth algorithm to mine frequent items in a frequent mode; the frequent item is a transaction contained in the session in the frequent mode;
if there is an intersection between the obtained transaction sequence and the frequent item in the frequent pattern, the frequent pattern is a matching pattern of the current real-time transaction sequence.
6. The power grid scheduling system page recommendation method based on pattern mining and correlation analysis of claim 5, wherein frequent pattern mining is performed by adopting a neighbor propagation clustering algorithm, and the method comprises the following steps:
establishing a session vector for the two sessions according to a 0-0 matching principle, calculating session similarity, and establishing a session similarity matrix;
clustering the sessions by adopting a neighbor propagation clustering algorithm, wherein each class corresponds to a frequent mode;
and determining the optimal clustering result by taking the contour coefficient as a standard for evaluating the clustering effect and adjusting the attenuation coefficient.
7. The method for recommending a page of a power grid scheduling system based on pattern mining and correlation analysis of claim 6, wherein the establishing of session vectors for two sessions according to the 0-0 matching principle comprises:
taking a transaction union set of the session 1 and the session 2 to form a transaction vector;
representing elements in the session vector by the number of times a transaction in the transaction vector occurs in the session;
calculating the conversation similarity as follows:
where s (x, y) represents the similarity of session x and session y.
8. The grid scheduling system page recommendation method based on pattern mining and correlation analysis as claimed in claim 6, wherein in the clustering process, an attenuation coefficient λ is introduced:
rt+1(i,k)←(1-λ)rt+1(i,k)+λrt(i,k)
at+1(i,k)←(1-λ)at+1(i,k)+λat(i,k)
where r (i, k) represents an element in the attraction matrix, a (i, k) is an element in the attribution matrix, i represents the ith session, k represents the kth session, and subscript t represents the tth iteration.
9. The method for recommending a page of a power grid scheduling system based on pattern mining and correlation analysis according to claim 1, wherein the remaining transactions except for the matching items in the frequent pattern having a pattern matching relationship with the real-time transaction sequence are obtained as recommended transactions, and further comprising:
calculating confidence degrees of the recommended transactions one by one, and selecting the transactions N before the confidence degree as recommended contents;
the confidence coefficient is as follows: probability of simultaneous occurrence of Y transactions on the premise of occurrence of X transactions; the calculation method is the ratio between the number of transactions containing X and Y transactions and the number of transactions containing X.
10. The power grid scheduling system page recommendation method based on pattern mining and correlation analysis of claim 9, wherein when the recommended transaction is an independent page, the independent page is given as a recommendation result; when the recommended transaction is a page class, a specific instance page of the page class needs to be given as a recommendation result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910812889.6A CN110599122A (en) | 2019-08-30 | 2019-08-30 | Power grid dispatching system page recommendation method based on pattern mining and correlation analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910812889.6A CN110599122A (en) | 2019-08-30 | 2019-08-30 | Power grid dispatching system page recommendation method based on pattern mining and correlation analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110599122A true CN110599122A (en) | 2019-12-20 |
Family
ID=68856797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910812889.6A Pending CN110599122A (en) | 2019-08-30 | 2019-08-30 | Power grid dispatching system page recommendation method based on pattern mining and correlation analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110599122A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112231598A (en) * | 2020-08-31 | 2021-01-15 | 咪咕文化科技有限公司 | Webpage path navigation method and device, electronic equipment and storage medium |
| CN113868544A (en) * | 2021-12-03 | 2021-12-31 | 杭银消费金融股份有限公司 | Intelligent service file processing method and service server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133370A (en) * | 2017-06-19 | 2017-09-05 | 南京邮电大学 | A kind of label recommendation method based on correlation rule |
| CN107451911A (en) * | 2017-07-19 | 2017-12-08 | 唐周屹 | A kind of method and system that real-time visual information is provided based on financial pipelined data |
| CN108334646A (en) * | 2018-04-11 | 2018-07-27 | 焦点科技股份有限公司 | A kind of link structure optimization method based on frequent browsing sequence |
-
2019
- 2019-08-30 CN CN201910812889.6A patent/CN110599122A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133370A (en) * | 2017-06-19 | 2017-09-05 | 南京邮电大学 | A kind of label recommendation method based on correlation rule |
| CN107451911A (en) * | 2017-07-19 | 2017-12-08 | 唐周屹 | A kind of method and system that real-time visual information is provided based on financial pipelined data |
| CN108334646A (en) * | 2018-04-11 | 2018-07-27 | 焦点科技股份有限公司 | A kind of link structure optimization method based on frequent browsing sequence |
Non-Patent Citations (3)
| Title |
|---|
| 朱鲲鹏: ""基于 Web 日志挖掘的智能信息检索研究"", 《中国博士学位论文全文数据库 信息科技辑》 * |
| 李中良: ""基于Web日志挖掘和关联规则的个性化推荐系统模型研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 班秋成: ""基于Hadoop的Web日志存储和分析系统的研究与实现"", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112231598A (en) * | 2020-08-31 | 2021-01-15 | 咪咕文化科技有限公司 | Webpage path navigation method and device, electronic equipment and storage medium |
| CN113868544A (en) * | 2021-12-03 | 2021-12-31 | 杭银消费金融股份有限公司 | Intelligent service file processing method and service server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bazan et al. | Dynamic reducts as a tool for extracting laws from decisions tables | |
| Frederickson | A data structure for dynamically maintaining rooted trees | |
| JP5092165B2 (en) | Data construction method and system | |
| US8135667B2 (en) | System, method, and computer-readable medium that facilitate in-database analytics with supervised data discretization | |
| CN111489201A (en) | Method, device and storage medium for analyzing customer value | |
| Liu et al. | Efficient mining of large maximal bicliques | |
| CN108229578B (en) | Image data target identification method based on three layers of data, information and knowledge map framework | |
| CN110389950B (en) | Rapid running big data cleaning method | |
| CN110287382B (en) | Mining method for association rule of battery production data | |
| CN110599122A (en) | Power grid dispatching system page recommendation method based on pattern mining and correlation analysis | |
| Ghesmoune et al. | Clustering over data streams based on growing neural gas | |
| Aronis et al. | Increasing the Efficiency of Data Mining Algorithms with Breadth-First Marker Propagation. | |
| CN111292201A (en) | Method for pushing field operation and maintenance information of power communication network based on Apriori and RETE | |
| CN108664548B (en) | Network access behavior characteristic group dynamic mining method and system under degradation condition | |
| CN116341059A (en) | Tunnel intelligent design method based on similarity | |
| CN100403308C (en) | SQL load mining-based automatic design method for physical database | |
| Li et al. | An efficient algorithm for mining high utility quantitative itemsets | |
| Miranda et al. | RegK-Means: A clustering algorithm using spatial contiguity constraints for regionalization problems | |
| CN109976271B (en) | Method for calculating information structure order degree by using information representation method | |
| Fu et al. | A specialized binary tree for financial time series representation | |
| Sun et al. | Continuous monitoring of maximum clique over dynamic graphs | |
| Babu et al. | Performance Analysis on Advances in Frequent Pattern Growth Algorithm | |
| CN106407296B (en) | Partial sweep correlation rule computer data analysis method based on anticipation screening | |
| CN112329432A (en) | Power distribution network voltage out-of-limit problem correlation analysis method based on improved Apriori | |
| Dubey et al. | Comparative Study of Frequent Item Set in Data Mining |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191220 |