Dispatcher operation recommendation method, recommendation device and recommendation systemTechnical Field
The invention relates to a dispatcher operation recommendation method, a recommendation device and a recommendation system, and belongs to the technical field of power grid dispatching system automation.
Background
With the continuous expansion of the power grid scale and the continuous upgrading of the dispatching system, the dispatching operation day is complex, fine and standard. For a typical scheduling working scene, (1) a scheduler needs to perform a plurality of fixed continuous operation actions on a specific page so as to finish a scheduling service, and (2) the scheduler needs to skip and browse layer by layer to analyze a plurality of fixed system pages, and reasonably operate by combining the data conditions of all devices and areas. In the above scenario, the operation mode of the dispatcher is fixed, the operation content is tedious, the efficiency is low, and the interaction cost with the dispatching system is too high, so an intelligent dispatching auxiliary scheme is needed to improve the work efficiency of the dispatcher, reduce the workload and reduce the work cost.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a dispatcher operation recommendation method, a recommendation device and a recommendation system, which can simplify the operation flow of a dispatcher and improve the work efficiency of the dispatcher and the intellectualization of a power grid dispatching system. In order to achieve the above purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the present invention provides a dispatcher operation recommendation method, including:
Acquiring an operation action of a dispatcher;
Judging a recommended mode of an operation action of a dispatcher to obtain a task recommended mode, wherein the task recommended mode comprises a real-time task recommended mode and a timing task recommended mode;
and determining corresponding operation recommendation according to the task recommendation mode.
With reference to the first aspect, further, according to the task recommendation mode, determining a corresponding operation recommendation includes:
converting the operation actions of the dispatcher into a real-time operation sequence in response to the recommended mode for the real-time task;
obtaining a session transaction sequence according to the real-time operation sequence;
Searching the FP_tree data structure to calculate the confidence level of the operation transaction in the first operation transaction set one by one, and selecting N operation transactions with the front N operation recommendations according to the confidence level;
responding that the session transaction sequence does not have intersection with the frequent pattern set of the dispatcher and has intersection with the frequent pattern sets of other dispatchers with the same attribute, and generating a second alternative operation transaction set according to the session transaction sequence and the frequent pattern set with intersection;
And in response to no intersection exists between the session transaction sequence and the dispatcher and other frequent pattern sets of the dispatcher with the same attribute, counting the historical operation transactions of the dispatcher to obtain a historical operation transaction set, searching the FP_tree data structure to calculate the confidence of the operation transactions in the historical operation transaction set one by one, and selecting the historical operation transactions with the N names before the confidence as operation recommendation.
With reference to the first aspect, further including updating the fp_tree data structure, including:
judging whether the operation transaction corresponding to the operation recommendation has nodes in the corresponding hierarchy of the pre-constructed FP_tree data structure, if so, adding 1 to the corresponding node count of the pre-constructed FP_tree data structure, and if not, adding a new operation action node to the corresponding hierarchy of the pre-constructed FP_tree data structure.
With reference to the first aspect, further, the method for constructing the frequent pattern set includes:
Traversing operation transactions of a dispatcher, calculating the support degree of each operation transaction, and creating a head pointer table according to the support degree of each operation transaction;
traversing operation sessions of a dispatcher, and screening intersections of each operation session and operation transactions in a head pointer table;
Sorting according to the occurrence frequency of each operation transaction in the intersection obtained by screening;
based on the FP_tree data structure and the support degree, generating a frequent pattern set of the dispatcher according to the sequencing result.
With reference to the first aspect, further, according to the task recommendation mode, determining a corresponding operation recommendation includes:
Dividing task periods of continuous fixed operation of the dispatcher in response to recommending a mode for the timed task;
And matching operation transactions for each task period from the historical operation transaction set of the dispatcher as operation recommendation.
With reference to the first aspect, further, the construction of the historical operation transaction set of the dispatcher includes:
and carrying out mode mining on the operation actions of the dispatcher in each period, and counting mining results to obtain a historical operation transaction set.
With reference to the first aspect, further, the determining a recommended mode of the dispatcher operating action includes:
tracking the operation action of the dispatcher according to the acquired operation action of the dispatcher;
If the operation action of the dispatcher is continuous fixed operation in a fixed period, the recommended mode of the operation action is a timed task recommended mode, otherwise, the recommended mode is a real-time task recommended mode.
With reference to the first aspect, further, the acquiring a dispatcher operation includes:
tracking the real-time operation action of the dispatcher by adopting an internal point burying mode to obtain real-time data of the operation action of the dispatcher;
and (3) retrieving a log file of the operation action of the dispatcher, and extracting information by a method of presetting a label to obtain the history data of the operation action of the dispatcher.
In a second aspect, the present invention provides a dispatcher-operated recommendation device, comprising:
The acquisition module is used for acquiring the operation actions of the dispatcher;
the judging module is used for judging the recommending mode of the operation action of the dispatcher to obtain a task recommending mode, wherein the task recommending mode comprises a real-time task recommending mode and a timing task recommending mode;
and the recommending module is used for determining corresponding operation recommendation according to the task recommending mode.
In a third aspect, the present invention provides a dispatcher-operated recommendation system, comprising:
The server side obtains recommendation data generated by operating the service device by the dispatcher and sends the recommendation data to the client side through a message bus, the client side displays operation action information and picture retrieval information, and the dispatcher executes recommendation operation or page opening operation through the client side;
The server side obtains statistical information of dispatcher related operations and page browsing frequency generated by the dispatcher operating service device, sends the statistical information to the client side through a message bus, the client side displays the related operation actions and the page browsing information and frequency, and the dispatcher executes recommendation operations or page opening operations through the client side;
The timing task module is used for identifying the time period task, sending the time period task to the client through the message bus, arranging the task from the time dimension by the client in a list mode, displaying the time period task, and executing the time period task by a dispatcher through the client;
And the mode library generating and updating module is used for generating a dispatcher operation behavior mode library by the server, comparing the generated mode with all modes in the mode library by the client, updating the mode library and then displaying the mode library if the mode does not exist in the mode library, otherwise, directly displaying the mode library, and checking the operation mode information in the mode library by the dispatcher through the client.
Compared with the prior art, the dispatcher operation recommendation method, the recommendation device and the recommendation system provided by the embodiment of the invention have the beneficial effects that:
the method comprises the steps of obtaining operation actions of a dispatcher, judging a recommended mode of the operation actions of the dispatcher to obtain a task recommended mode, wherein the task recommended mode comprises a real-time task recommended mode and a timing task recommended mode, and determining corresponding operation recommendation according to the task recommended mode. The invention is based on the historical operation record of the dispatcher in the dispatching system, combines the data analysis and the machine learning algorithm, effectively mines the operation behavior mode of the dispatcher, combines the modes of associated recommendation, statistical recommendation, collaborative recommendation and the like to realize the accurate recommendation of the dispatcher operation, can reduce the operation flow of the dispatcher on the page of the power grid dispatching system, provide accurate operation actions for the dispatcher, improve the interaction mode of the dispatcher and the power grid dispatching system, lighten the workload of the dispatcher, improve the operation convenience of the dispatcher and enhance the intellectualization of the power grid dispatching system.
Drawings
FIG. 1 is a flowchart of a dispatcher operation recommendation method provided in embodiment 1 of the present invention;
FIG. 2 is a flowchart of a method for recommending scheduler operations to construct frequent pattern sets according to embodiment 1 of the present invention;
fig. 3 is a frame diagram of a dispatcher-operated recommendation device according to embodiment 3 of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for more clearly illustrating the technical aspects of the present invention, and are not intended to limit the scope of the present invention.
Embodiment one:
as shown in fig. 1, an embodiment of the present invention provides a dispatcher operation recommendation method, including:
Acquiring an operation action of a dispatcher;
Judging a recommended mode of an operation action of a dispatcher to obtain a task recommended mode, wherein the task recommended mode comprises a real-time task recommended mode and a timing task recommended mode;
and determining corresponding operation recommendation according to the task recommendation mode.
And step 1, acquiring operation actions of a dispatcher.
Tracking the real-time operation action of the dispatcher by adopting an internal point burying mode to obtain real-time data of the operation action of the dispatcher;
and (3) retrieving a log file of the operation action of the dispatcher, and extracting information by a method of presetting a label to obtain the history data of the operation action of the dispatcher.
And 2, judging a recommended mode of the operation action of the dispatcher to obtain a task recommended mode.
Tracking the operation action of the dispatcher according to the acquired operation action of the dispatcher;
If the operation action of the dispatcher is continuous fixed operation in a fixed period, the recommended mode of the operation action is a timed task recommended mode, otherwise, the recommended mode is a real-time task recommended mode.
Converting the operation actions of the dispatcher into a real-time operation sequence in response to the recommended mode for the real-time task;
obtaining a session transaction sequence according to the real-time operation sequence;
searching an FP_tree data structure to calculate the confidence level of operation transactions in the first operation transaction set one by one, and selecting N operation transactions with the front N confidence levels as operation recommendations;
The collaborative recommendation is that in response to the fact that the session transaction sequence does not have an intersection with the frequent pattern set of the dispatcher and has an intersection with the frequent pattern sets of other dispatchers with the same attribute, a second alternative operation transaction set is generated according to the session transaction sequence and the frequent pattern set with the intersection;
And (3) counting and recommending, namely counting the historical operation transactions of the dispatcher to obtain a historical operation transaction set in response to no intersection between the session transaction sequence and the frequent pattern set of the dispatcher and other frequent pattern sets of the dispatcher with the same attribute, searching the FP_tree data structure to calculate the confidence of the operation transactions in the historical operation transaction set one by one, and selecting the historical operation transactions with the N front confidence as operation recommendation.
In some embodiments, the co-attribute dispatcher may be a dispatcher with a scene similar to a post job.
Updating the fp_tree data structure, comprising:
judging whether the operation transaction corresponding to the operation recommendation has nodes in the corresponding hierarchy of the pre-constructed FP_tree data structure, if so, adding 1 to the corresponding node count of the pre-constructed FP_tree data structure, and if not, adding a new operation action node to the corresponding hierarchy of the pre-constructed FP_tree data structure.
The construction method of the frequent pattern set comprises the following steps:
Traversing operation transactions of a dispatcher, calculating the support degree of each operation transaction, and creating a head pointer table according to the support degree of each operation transaction;
traversing operation sessions of a dispatcher, and screening intersections of each operation session and operation transactions in a head pointer table;
Sorting according to the occurrence frequency of each operation transaction in the intersection obtained by screening;
based on the FP_tree data structure and the support degree, generating a frequent pattern set of the dispatcher according to the sequencing result.
Dividing task periods of continuous fixed operation of the dispatcher in response to recommending a mode for the timed task;
And matching operation transactions for each task period from the historical operation transaction set of the dispatcher as operation recommendation.
The construction of the historical operation transaction set of the dispatcher comprises the following steps:
and carrying out mode mining on the operation actions of the dispatcher in each period, and counting mining results to obtain a historical operation transaction set.
The recommendation method further includes pattern analysis. Pattern analysis includes both critical transaction page analysis and pattern frequency analysis. The key business pages are business pages frequently browsed by regulatory personnel in the same type of session and are used for finding main pages in an operation mode. The pattern frequency is the proportion of the number of times of pattern occurrence to the number of days in the statistical time period, and is used for finding the rule of the pattern in the time dimension.
The recommendation method further includes session division. Session partitioning includes session identification and transaction identification.
And carrying out session identification by using 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 session is stopped, and a new session is generated. Wherein the time interval of two consecutive pages is defined as the page dwell time. Typically the session identification threshold is set to 600s.
A transaction refers to a sequence of pages within a session that are browsed to complete a single step in the task, with one or more transactions in a session. Transaction identification is performed herein using a reference length method. The reference length method divides pages into content pages and navigation pages. The content page is a page with access purpose and interested by the regulating personnel, and the navigation page is a page with jump property of the regulating personnel in the browsing process. The page view records are sequentially read and once a transition from a navigation page to a content page is found, the boundaries of the transaction are then determined.
Embodiment two:
the embodiment of the invention provides a dispatcher operation recommendation device, which comprises:
The acquisition module is used for acquiring the operation actions of the dispatcher;
the judging module is used for judging the recommending mode of the operation action of the dispatcher to obtain a task recommending mode, wherein the task recommending mode comprises a real-time task recommending mode and a timing task recommending mode;
and the recommending module is used for determining corresponding operation recommendation according to the task recommending mode.
Embodiment III:
As shown in fig. 3, an embodiment of the present invention provides a dispatcher-operated recommendation system, including:
the server side obtains recommendation data generated by a dispatcher operating the recommendation device, the recommendation data are sent to the client side through a message bus, the client side displays operation action information and picture retrieval information, and the dispatcher executes recommendation operation or page opening operation through the client side;
The server side obtains statistical information of dispatcher related operations and page browsing frequency generated by the dispatcher operation recommendation device, the statistical information is sent to the client side through a message bus, the client side displays the related operation actions and the page browsing information and frequency, and the dispatcher executes recommendation operations or page opening operations through the client side;
The timing task module is used for identifying the time period task, sending the time period task to the client through the message bus, arranging the task from the time dimension by the client in a list mode, displaying the time period task, and executing the time period task by a dispatcher through the client;
And the mode library generating and updating module is used for generating a dispatcher operation behavior mode library by the server, comparing the generated mode with all modes in the mode library by the client, updating the mode library and then displaying the mode library if the mode does not exist in the mode library, otherwise, directly displaying the mode library, and checking the operation mode information in the mode library by the dispatcher through the client.
The foregoing is merely a preferred embodiment of the present invention, and it should be noted that modifications and variations could be made by those skilled in the art without departing from the technical principles of the present invention, and such modifications and variations should also be regarded as being within the scope of the invention.