CN115953123A - Method, device and equipment for generating robot automation flow and storage medium - Google Patents
Method, device and equipment for generating robot automation flow and storage medium Download PDFInfo
- Publication number
- CN115953123A CN115953123A CN202211630249.1A CN202211630249A CN115953123A CN 115953123 A CN115953123 A CN 115953123A CN 202211630249 A CN202211630249 A CN 202211630249A CN 115953123 A CN115953123 A CN 115953123A
- Authority
- CN
- China
- Prior art keywords
- behavior
- sequence
- target
- sequences
- events
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The embodiment of the application provides a method, a device, equipment and a storage medium for generating a robot automation flow, wherein the method for generating the robot automation flow comprises the following steps: acquiring operation data of a target user for operating the electronic equipment; dividing the operation data into a plurality of behavior sequences according to a preset time interval; dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence; extracting a public sequence of repeated work from the behavior sequences of all categories; and generating a robot automation flow according to the public sequence. According to the embodiment of the application, the behavior sequences can be automatically divided, the public sequences are extracted from the behavior sequences of various categories and used for generating the robot automation process, the automation degree of process design is improved, the identification accuracy and the usability are higher, and the adaptability to the redundant operation of a user is high and the tolerance is higher.
Description
Technical Field
The application belongs to the technical field of artificial intelligence, and particularly relates to a method, a device, equipment and a storage medium for generating a robot automation process.
Background
Robot Process Automation (RPA) automates user operations by simulating user operations in a software system based on techniques such as interface recognition, browser operations, and Artificial Intelligence (AI). The enterprises use the RPA capability to extract the activities with high repeatability and clear business logic in the work into the RPA flow to realize automation, and the aims of IT people changing, cost reduction and efficiency improvement are achieved by utilizing the characteristics of high accuracy and 7 × 24 hour operation of the RPA.
At present, before designing an RPA process, a user and an implementer are required to perform manual demand analysis on the working content of the user, the RPA process which is high in repeatability and can be automated is screened out from the working of the user, and the RPA process is designed from zero according to the working content of the RPA process. In the two stages of demand analysis and flow design, manual analysis and design are long in time consumption and low in efficiency due to high working complexity of users, and the purposes of cost reduction and efficiency improvement of the RPA are violated.
Disclosure of Invention
The embodiment of the application provides a method, a device, equipment and a storage medium for generating a robot automation flow, and can improve the extraction efficiency of the robot automation flow.
In a first aspect, an embodiment of the present application provides a method for generating a robot automation flow, where the method for generating a robot automation flow includes: acquiring operation data of a target user for operating the electronic equipment; dividing the operation data into a plurality of behavior sequences according to a preset time interval; dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence; extracting a public sequence of repeated work from the behavior sequences of all categories; and generating a robot automation flow according to the public sequence.
According to an embodiment of the first aspect of the present application, acquiring operation data of a target user operating an electronic device specifically includes: acquiring an operation event, target window information, target element information and time of the operation event of an input device of the electronic equipment; generating an operation log according to the operation event, the target window information, the target element information and the time of the operation event; and obtaining operation data according to the at least one operation log.
According to any of the previous embodiments of the first aspect of the present application, the target window information comprises browser window information and desktop application window information; the method for acquiring the operation event, the target window information, the target element information and the time of the operation event of the input device of the electronic equipment specifically comprises the following steps: when an operation event is detected, acquiring target window information; identifying the type of the target window information; when the target window information is browser window information, acquiring target element information corresponding to the browser window information through a browser plug-in and a target engine component; and when the target window information is the desktop application program window information, acquiring target element information corresponding to the desktop application program window information through the application element identification unit.
According to any of the foregoing embodiments of the first aspect of the present application, the dividing, based on a clustering algorithm, a plurality of behavior sequences into a plurality of behavior sequences of different categories specifically includes: performing word segmentation processing on the plurality of behavior sequences based on a character string matching algorithm of a dictionary and/or a pre-trained machine learning model to obtain a plurality of behavior sequences after word segmentation processing; converting target window information and target element information in the behavior sequences after word segmentation into target vectors to obtain behavior sequence vector sets corresponding to the behavior sequences; and dividing each behavior sequence vector set into a plurality of behavior sequences of different categories based on a clustering algorithm.
According to any one of the foregoing embodiments of the first aspect of the present application, before performing word segmentation processing on a plurality of behavior sequences, the method for dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm further includes: acquiring the frequency or frequency of each behavior in a plurality of behavior sequences; and when the frequency or the frequency count is smaller than a first preset threshold value, deleting the behavior corresponding to the frequency or the frequency count smaller than the first preset threshold value from the behavior sequence.
According to any one of the foregoing embodiments of the first aspect of the present application, after performing word segmentation processing on a plurality of behavior sequences, dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, specifically further includes: and deleting the words of which the frequency in the behavior sequences after word segmentation is greater than a second preset threshold value.
According to any one of the preceding embodiments of the first aspect of the present application, the sequence of actions comprises a plurality of action events; extracting a common sequence of repeated work from the behavior sequences of each category, which specifically comprises the following steps: comparing each behavior event in the behavior sequence of the same category based on a preset comparison strategy, and screening out target behavior events which are consistent with each other; combining the target behavioral events into a common sequence.
According to any one of the previous embodiments of the first aspect of the present application, the behavior event includes an operation event of an input device of the electronic device, target window information, target element information, and a time of the operation event; based on a preset comparison strategy, comparing all behaviors in the behavior sequence of the same category, and screening out target behaviors which are consistent with each other, specifically comprising the following steps: if the types of the operation events in any two behavior events are inconsistent, judging that any two behavior events are inconsistent; if the types of the operation events in any two behavior events are consistent and the event values corresponding to the operation events are inconsistent, judging that any two behavior events are inconsistent; if the target window information in any two behavior events is inconsistent, judging that any two behavior events are inconsistent; calculating the similarity between target element information in any two behavior events based on a natural language algorithm; and if the similarity is smaller than a third preset threshold, judging that any two behavior events are inconsistent.
According to any of the foregoing embodiments of the first aspect of the present application, after extracting the common sequence of repetitive work from the behavior sequences of the respective categories, the method for generating the robot automation process further includes: and generating a requirement definition document according to the operation events, the target window information and the target element information in the public sequence, wherein the requirement definition document is marked with specific work steps and time consumption of manual execution.
According to any one of the preceding embodiments of the first aspect of the present application, the generating of the robot automation process according to the common sequence specifically includes: and defining a document according to the public sequence and the requirement, and generating an RPA flow template based on the RPA plug-in of the robot automation flow.
In a second aspect, an embodiment of the present application provides an apparatus for generating a robot automation flow, where the apparatus for generating a robot automation flow includes: the acquisition module is used for acquiring operation data of a target user for operating the electronic equipment; the dividing module is used for dividing the operation data into a plurality of behavior sequences according to a preset time interval; the clustering module is used for dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence; the extraction module is used for extracting a public sequence of repeated work from the behavior sequences of all categories; and the generating module is used for generating the robot automation flow according to the public sequence.
In a third aspect, an embodiment of the present application provides an electronic device, where the electronic device includes: a processor, a memory and a computer program stored on the memory and executable on the processor, the computer program, when executed by the processor, implementing the steps of the method of generating a robot automation flow as provided by the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the method for generating a robot automation process as provided in the first aspect.
The method, the device, the equipment and the storage medium for generating the robot automation process acquire operation data of a target user for operating the electronic equipment; dividing the operation data into a plurality of behavior sequences according to a preset time interval; dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence; extracting a public sequence of repeated work from the behavior sequences of all categories; and generating a robot automation flow according to the public sequence. The acquired operation data are automatically divided based on the preset time interval, the problem that the behavior sequence can only be divided by requiring a user to specify the starting action and the ending action of the behavior sequence in the prior art is solved, the automation degree of flow design is improved, the behavior sequence is divided into categories based on the clustering algorithm, a public sequence is extracted from the behavior sequence of each category, and compared with the prior art that the similarity of operation events is directly compared, the method has higher identification accuracy and usability, and has strong adaptability and higher tolerance on the redundant operation of the user.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for generating a robot automation flow according to an embodiment of the present disclosure;
fig. 2 is a schematic flow chart diagram of another method for generating a robot automation flow according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of behavior alignment provided by an embodiment of the present application;
fig. 4 is a schematic structural diagram of a generating apparatus of a robot automation flow provided in an embodiment of the present application;
fig. 5 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application.
Detailed Description
Features and exemplary embodiments of various aspects of the present application will be described in detail below, and in order to make objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are intended to be illustrative only and are not intended to be limiting. It will be apparent to one skilled in the art that the present application may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present application by illustrating examples thereof.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising 8230; \8230;" comprises 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
It should be understood that the term "and/or" as used herein is merely one type of association that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" herein generally indicates that the former and latter related objects are in an "or" relationship.
It will be apparent to those skilled in the art that various modifications and variations can be made in the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application cover the modifications and variations of this application provided they come within the scope of the corresponding claims (the claimed technology) and their equivalents. It should be noted that the embodiments provided in the embodiments of the present application can be combined with each other without contradiction.
Before explaining the technical solutions provided by the embodiments of the present application, in order to facilitate understanding of the embodiments of the present application, the present application first specifically explains the problems existing in the related art:
as described above, the inventors of the present application have found that the following problems exist in the design of the RPA procedure in the related art: firstly, the automation degree is low, when the operation data of a user is collected, the user is required to inform the data collector when the task starts and ends each time, so that the segmentation of the behavior sequence is realized, and the automatic segmentation of the operation sequence cannot be realized; secondly, the robustness of the behavior similarity detection function is low, only the Natural Language Processing (NLP) result is used as the basis for behavior similarity judgment, the Processing capability of the similar elements in different windows and the input conditions of different key values in the same operation are poor, and the misjudgment rate is high; thirdly, the robustness of the behavior sequence extraction function is low, public sequence extraction needs to be achieved through a sliding window function, the user is required to operate completely consistently when repeated activities are carried out, the adaptability to redundant operation is poor, and the requirement on the operation normalization of the user is high.
In view of the above research findings of the inventors, embodiments of the present application provide a method, an apparatus, a device, and a storage medium for generating a robot automation flow, which can solve the technical problems of low automation degree, low robustness of a behavior similarity detection function, and low robustness of a behavior sequence extraction function in the related art.
First, a method for generating a robot automation flow provided in the embodiment of the present application is described below.
Fig. 1 is a schematic flow chart of a method for generating a robot automation flow according to an embodiment of the present disclosure. As shown in fig. 1, the method may include the steps of:
s101, obtaining operation data of a target user for operating the electronic equipment.
When a target user performs various operations on the electronic equipment, various operation data of the target user during the operations are obtained according to operation events, target window information, target element information and operation event time of the target user on an input device of the electronic equipment.
S102, dividing the operation data into a plurality of behavior sequences according to a preset time interval.
The time interval for dividing the behavior sequence is preset according to the time of the target user operation event, the operation data is divided into a plurality of behavior sequences according to the preset time interval, manual operation is not needed, and each behavior sequence is an operation flow of one work.
S103, dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence.
And clustering the plurality of divided behavior sequences based on a clustering algorithm, and further dividing a plurality of behavior sequences of different categories, wherein the behavior sequence of each category comprises at least one behavior sequence, and the behavior sequences in the same category are the work tasks with the same content.
And S104, extracting a common sequence of repeated work from the behavior sequences of the various categories.
And respectively carrying out similarity comparison on the behavior sequences in the same category, extracting a repeatedly working public sequence in the same category, and further obtaining the repeatedly working public sequence in each category.
And S105, generating a robot automation flow according to the public sequence.
And generating a robot automation flow according to the extracted common sequence of repeated work in each category, and the operation event, the target window information, the target element information and the time of the operation event when the target user operates the input device of the electronic equipment.
According to the method for generating the robot automation process, operation data of a target user for operating the electronic equipment are obtained; dividing the operation data into a plurality of behavior sequences according to a preset time interval; dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence; extracting a public sequence of repeated work from the behavior sequences of all categories; and generating a robot automation flow according to the public sequence. The acquired operation data are automatically divided based on the preset time interval, the problem that the behavior sequence can only be divided by requiring a user to specify the starting action and the ending action of the behavior sequence in the prior art is solved, the automation degree of flow design is improved, the behavior sequence is divided into categories based on a clustering algorithm, a public sequence is extracted from the behavior sequence of each category, and compared with the prior art that the similarity of operation events is directly compared, the method has higher identification accuracy and usability, and has stronger adaptability to the redundant operation of the user and higher tolerance.
In some embodiments, the obtaining operation data of the target user operating the electronic device specifically includes: acquiring an operation event, target window information, target element information and time of the operation event of an input device of the electronic equipment; generating an operation log according to the operation event, the target window information, the target element information and the time of the operation event; and obtaining operation data according to the at least one operation log.
Illustratively, the acquired operation events of the input device of the electronic device may be operation events of a mouse and a keyboard of the computer by the target user, such as operation events of left-clicking a mouse, right-clicking a mouse, single-time pressing, combined pressing, continuous pressing and the like by the target user. The corresponding operation event type may be a mouse event or a keyboard event, the operation event value is an operation click key value, and the operation event type, the operation event value and the time of the operation event are recorded in an operation log, for example, a mouse event-left click, a mouse event-right click, a keyboard event-a, and the like. In the actual operation activities, because the target user has operation behaviors of single key pressing, combined key pressing, continuous key pressing and the like on the keyboard, in order to improve the accuracy of subsequent public sequence extraction, the keyboard events need to be processed, and specifically, the keyboard events may include, but are not limited to, the operation events of combined key pressing and continuous key pressing of the target user are integrated and recorded in the operation log.
The target window may include, but is not limited to, a browser window and a desktop application window. When a target user performs mouse operation or keyboard operation, target window information, that is, browser window information or desktop application window information, is acquired based on the operation of the target user and recorded in an operation log.
The target element information may include, but is not limited to, target element information in a browser window and target element information in a desktop application window, where the target element information in the browser window may be extensible markup Language Path Language (XPath), JSPath, etc. information that a target user operates a target element. The target element information may serve as an identifier of the target element, and the target element information needs to be recorded in the operation log.
According to at least one operation log, various operation data of the target user for operating the electronic equipment, which are recorded in the operation log, can be acquired.
In some embodiments, the target window information includes browser window information and desktop application window information; the method for acquiring the operation event, the target window information, the target element information and the time of the operation event of the input device of the electronic equipment specifically comprises the following steps: when an operation event is detected, acquiring target window information; identifying the type of the target window information; when the target window information is browser window information, acquiring target element information corresponding to the browser window information through a browser plug-in and a target engine component; and when the target window information is the desktop application program window information, acquiring the target element information corresponding to the desktop application program window information through the application element identification unit.
Illustratively, when an operation event such as a mouse operation or a keyboard operation performed by a target user is detected, target window information is acquired based on the operation event, and the type of the target window information is determined. When the target window information is browser window information, the communication with the browser is realized through the browser plug-in and the target engine component, and then the target element information in the browser window, namely the target element information corresponding to the browser window information, is obtained. Where browsers may include, but are not limited to, chrome browsers and FireFox browsers, target engine components may include, but are not limited to, MSHTML components. And when the target window information is the desktop application program window information, acquiring target element information in the desktop application program window through the application identifier, namely the target element information corresponding to the desktop application program window information.
Fig. 2 is a schematic flow chart of another method for generating a robot automation flow according to an embodiment of the present disclosure. As shown in fig. 2, according to some embodiments of the present application, optionally, S103 divides the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, where the behavior sequence of each category includes at least one behavior sequence, and may further include the following steps S201 to S203.
S201, performing word segmentation processing on the plurality of behavior sequences based on a dictionary-based character string matching algorithm and/or a pre-trained machine learning model to obtain a plurality of behavior sequences after word segmentation processing.
When the behavior sequences are divided into a plurality of behavior sequences of different types based on a clustering algorithm, preprocessing needs to be performed on each behavior sequence, specifically including performing word segmentation processing on target window information and target element information of the behavior sequences. Word segmentation is an important basic preprocessing step in natural language processing, and the text is divided into smaller semantic unit words so as to be convenient for subsequent use and analysis. Different from English words which are naturally separated by spaces, chinese word segmentation is more difficult because Chinese text words have no space.
The chinese segmentation algorithm may include, but is not limited to, a dictionary-based string matching algorithm and a machine learning model-based segmentation algorithm. The dictionary-based string matching algorithm relies entirely on the matching results by building a sufficiently large dictionary base for matching strings. The algorithm is divided into a forward maximum matching method, a reverse maximum matching method and a bidirectional matching word-dividing method according to different priorities of character string identification sequences and sizes. The dictionary-based character string matching algorithm can achieve efficient word segmentation of linear time complexity through optimization of a data structure stored in a character string and a search algorithm, such as hash table storage and an index tree, but the word segmentation effect on new words and ambiguous words outside a dictionary is not ideal.
In addition, no matter a standard word segmenter, a Natural Language Processing (NLP) word segmenter or an index word segmenter, a target result word cannot be accurately segmented every time, and custom nouns such as "apple cell phone", "demand intelligent mining" and the like often appear in the current internet, and the word segmentation function has a defect of low accuracy rate in the face of a long sentence. In order to make up for the defect, a user needs to define a dictionary by user definition to record common words, so that the word segmentation device can realize high-accuracy word segmentation based on the user-defined dictionary. According to the embodiment of the application, the professional domain knowledge dictionary can be selected according to the target domain of the enterprise, so that more detailed word segmentation can be realized, for example, specific words in the financial domain are collected, and the financial domain dictionary is customized, so that the accuracy of behavior sequence classification is improved when an enterprise such as a bank, securities and the like generates a robot automation process.
The machine learning Model-based segmentation algorithm may include, but is not limited to, a Hidden Markov Model (HMM) based on a generative Model segmentation algorithm and a Conditional Random Field Model (CRF) based on a discriminant Model segmentation algorithm. The machine learning model does not perform word segmentation matching by taking words as units, but labels the words, models a hidden Markov model or a conditional random field model for a given labeled sequence Y by considering the relation between contexts, and predicts the probability that an observation sequence X appears behind the labeled sequence Y and forms words. The word segmentation algorithm based on the machine learning model does not depend on a dictionary and has better learning capacity for new words and ambiguous words.
And performing word segmentation processing on the plurality of behavior sequences based on any one method or combination of a dictionary character string matching algorithm and a pre-trained machine learning model to obtain a plurality of behavior sequences after word segmentation processing.
S202, converting target window information and target element information in the behavior sequences after word segmentation into target vectors to obtain behavior sequence vector sets corresponding to the behavior sequences.
Because the text content in the target window information and the target element information of each behavior in the behavior sequence cannot be directly recognized by the computer, the text needs to be vectorized and expressed on the premise of keeping the text characteristics as much as possible. The words are used as basic units containing specific meanings in natural language, and word vectors obtained by mapping the words to real number field vectors are the most commonly used text features in natural language processing. Such mapping methods generally fall into two categories: one is discrete representation and the other is distributed representation.
The discrete representation of a word vector is based on a Bag of Words (BOW), that is, a text is represented by a set of unordered Words that is, a text is regarded as an unordered set formed by a series of Words constituting the text, regardless of the semantic, the word order, and the structure of the text. Such as one-hot encoding and word frequency-inverse document frequency.
A distributed representation of a word vector is also called word embedding, i.e. the semantics of a word are determined by its context, words with similar context semantics are more likely to have similar semantics. Such as google-sourced tool Word2Vec that directly obtains low-dimensional Word vectors.
And converting the behavior sequences subjected to word segmentation processing into corresponding target vectors based on a discrete representation mapping method or a distributed representation mapping method to obtain a behavior sequence vector set which is used as the input of a subsequent clustering algorithm.
And S203, dividing each behavior sequence vector set into a plurality of behavior sequences of different categories based on a clustering algorithm.
Common clustering algorithms are mainly based on three ideas: the partition Method (Partitioning Method), the Hierarchical Method (Hierarchical Method) and the Density-based Method (Density-based Method). The Clustering algorithm employed in the embodiments of the present application may include, but is not limited to, K-Means Clustering algorithm, split-level word Clustering algorithm (divorce Clustering), and agglomerative-level Clustering algorithm.
The behavior sequence vector set is used as the input of a clustering algorithm, each behavior sequence vector set is divided into a plurality of behavior sequences of different categories based on the clustering algorithm, the behavior sequences of the same category have high similarity, the behavior sequences of different categories have as high dissimilarity as possible, and each category of behavior sequences respectively represents a repetitive work with automation potential.
In some embodiments, before performing word segmentation processing on the plurality of behavior sequences, the dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm further includes: acquiring the frequency or frequency of each behavior in a plurality of behavior sequences; and when the frequency or the frequency count is smaller than a first preset threshold value, deleting the behavior corresponding to the frequency or the frequency count smaller than the first preset threshold value from the behavior sequence.
Illustratively, when a target user operates the electronic device, redundant operations inevitably occur, which are not related to actual work content, but interfere with the clustering effect of the behavior sequence and reduce the extraction efficiency of the common sequence, so that redundant behaviors in the behavior sequence need to be removed. Comparing each behavior sequence in the behavior sequences to obtain the frequency or frequency of each behavior, regarding the behavior with the frequency or frequency smaller than a first preset threshold as a redundant behavior, and deleting the redundant behavior from the behavior sequences. The setting of the first preset threshold will determine the tolerance degree for the redundant behavior, the determination condition of the threshold may be frequency or frequency, the first preset threshold and the determination condition thereof may be specifically determined according to the actual application scenario, for example, when the first preset threshold of the frequency is set to 1, the redundant behavior that only appears in a single behavior sequence is deleted.
In some embodiments, after performing word segmentation processing on the plurality of behavior sequences, the method of dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm further includes: and deleting the words of which the frequency in the behavior sequences after word segmentation is greater than a second preset threshold value.
Illustratively, some words occur in natural language with a very high frequency, but without practical meaning, and such words are referred to as stop words in natural language processing. In order to avoid the interference of the words with the subsequent common sequences for extracting repetitive work, the stop words with the frequency larger than a second preset threshold value in the behavior sequences are deleted based on the stop word list after the word segmentation processing is carried out on the plurality of behavior sequences. The stop word list may be custom defined or may use an existing stop word list, which may include, but is not limited to, a Haugh stop word list and a Baidu stop word list.
In some embodiments, the sequence of behaviors includes a plurality of behavioral events; extracting a common sequence of repeated work from the behavior sequences of each category, which specifically comprises the following steps: comparing each behavior event in the behavior sequence of the same category based on a preset comparison strategy, and screening out target behavior events which are consistent with each other; combining the target behavior events into a common sequence.
Illustratively, the common sequence is the shortest behavior sequence capable of realizing the target task, and the extraction of the common sequence is generally divided into two parts of behavior consistency judgment and behavior sequence comparison. The behavior consistency judgment is to judge whether two behavior events in the same category of behavior sequences are consistent or not from three dimensions of the operation event, the target window information and the target element information based on a preset comparison strategy. The behavior sequence comparison is to compare a plurality of behavior sequences based on a preset comparison strategy, and extract a public sequence from the behavior sequences. Compared with the prior art of directly comparing the similarity of the operation events, the method has higher identification accuracy on the operation events, the target window information and the target element information.
Fig. 3 is a schematic flow chart of behavior alignment provided in the embodiment of the present application, as shown in fig. 3. Taking three-sequence comparison as an example, based on a preset comparison strategy, the longest common subsequence in the three behavior sequences of the sequence 301, the sequence 302 and the sequence 303 is obtained, wherein the subsequence may be discontinuous in each original sequence, so that the longest common subsequence 304 is obtained, and the rest behaviors are regarded as redundant behaviors to be discarded. Based on this, the extraction of a common sequence for all behavior sequences in the same category can be achieved. Compared with the prior art which requires the operation sequence to be completely consistent or the starting behaviors to be consistent, the method for extracting the longest public subsequence has stronger adaptability to the redundant operation of the user, higher tolerance and low requirement on the operation normative of the user, and does not need the user to specify the sequence starting behaviors or operate completely consistent when repeated activities are carried out.
Through behavior consistency judgment and behavior sequence comparison, a common sequence of all behavior sequences in the same category can be output, and the common sequence is used as a standard flow of the behavior sequences. The method is adopted to extract the public sequence for the behavior sequences of all categories, so that all the work flows of the target user are analyzed.
In some embodiments, the behavior event includes an operation event of an input device of the electronic device, target window information, target element information, and a time of the operation event; based on a preset comparison strategy, comparing all behaviors in the behavior sequence of the same category, and screening out target behaviors which are consistent with each other, specifically comprising the following steps: if the types of the operation events in any two behavior events are inconsistent, judging that any two behavior events are inconsistent; if the types of the operation events in any two behavior events are consistent and the event values corresponding to the operation events are inconsistent, judging that any two behavior events are inconsistent; if the target window information in any two behavior events is inconsistent, judging that any two behavior events are inconsistent; calculating the similarity between target element information in any two behavior events based on a natural language algorithm; and if the similarity is smaller than a third preset threshold, judging that any two behavior events are inconsistent.
Illustratively, based on a preset comparison strategy, whether two behavior events in a behavior sequence of the same category are consistent or not is judged from three dimensions of an operation event, target window information and target element information, behavior events which are consistent with each other are screened out to serve as target behaviors, multi-dimensional behavior consistency judgment can adapt to behavior similarity comparison operation under various conditions, and robustness of data analysis is enhanced.
If the types of the operation events in any two behavior events are inconsistent, such as a mouse event and a keyboard event, judging that any two behavior events are inconsistent; if the types of the operation events in any two behavior events are consistent, and the event values corresponding to the operation events are inconsistent, such as mouse event-left click and mouse event-right click, keyboard event-a and keyboard event-b, or the event values of the keyboard event belong to different keyboard combinations such as text input combination, special key value combination and the like, judging that any two behavior events are inconsistent; if the target window information in any two behavior events is inconsistent, such as a browser window and a desktop application program window, judging that any two behavior events are inconsistent; calculating the similarity between target element information in any two behavior events based on a natural language algorithm; and if the similarity is smaller than a third preset threshold, judging that any two behavior events are inconsistent.
In some embodiments, after extracting the common sequence of repetitive work from the behavior sequences of the respective categories, the method for generating a robot automation flow further includes: and generating a requirement definition document according to the operation events, the target window information and the target element information in the public sequence, wherein the requirement definition document is marked with specific work steps and time consumption of manual execution.
Illustratively, based on the common sequence of repeated work extracted from the behavior sequences of various categories, a requirement definition document is generated according to the operation events, the target window information and the target element information of various behaviors of the common sequence, and the specific steps of the work with high automation potential and the time consumed for manual execution are written in the requirement definition document and serve as the basis for the subsequent RPA requirement analysis.
In some embodiments, the generating the robotic automation process from the common sequence includes: and defining a document according to the public sequence and the requirement, and generating an RPA flow template based on the RPA plug-in of the robot automation flow.
Illustratively, RPA flow templates are automatically generated based on common sequence and requirement definition documents in conjunction with robotic automation flow RPA plug-ins. The RPA flow template can be directly used as a service flow to be applied to service automation, and can also be used as a development template to carry out flow optimization or secondary development so as to shorten the flow design period.
Based on the method for generating a robot automation flow provided in the foregoing embodiment, correspondingly, the application 5 also provides a specific implementation manner of a device for generating a robot automation flow. Please refer to the following examples
Examples are shown.
Referring first to fig. 4, a device 40 for generating a robot automation process provided in an embodiment of the present application includes the following modules:
an obtaining module 401, configured to obtain operation data of a target user operating an electronic device; a 0 dividing module 402, configured to divide the operation data into a plurality of rows according to a preset time interval
Is a sequence;
a clustering module 403, configured to divide the behavior sequences into behavior sequences of different categories based on a clustering algorithm, where a behavior sequence of each category includes at least one behavior sequence;
an extracting module 404, configured to extract a common 5 sequence of repetitive work from the behavior sequences of the respective categories;
a generating module 405 for generating a robot automation flow according to the common sequence.
The device for generating the robot automation process obtains operation data of a target user for operating the electronic equipment; dividing operation data into a plurality of line sequences according to a preset time interval
Columns; dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein each behavior sequence of category 0 comprises at least one behavior sequence; extracting from the behavior sequences of the various categories
A common sequence of repetitive work; and generating a robot automation flow according to the public sequence. The acquired operation data are automatically divided based on the preset time interval, so that the problem that the behavior sequence can be divided only by requiring a user to specify the starting and ending actions of the behavior sequence in the prior art is solved, and the method and the device for dividing the behavior sequence are provided
The automation degree of the process design is improved, the behavior sequences are classified based on a clustering algorithm, 5, a public sequence is extracted from the behavior sequences of each class, and compared with the prior art that the operation is directly compared
The similarity of events has higher identification accuracy and usability, and has strong adaptability and higher tolerance to the redundant operation of the user.
In some embodiments, the obtaining module 401 is specifically configured to: acquiring an operation event, target window information, target element information and time of the operation event of an input device of the electronic equipment; generating an operation log according to the operation event, the target window information, the target element information and the time of the operation event; and obtaining operation data according to the at least one operation log.
In some embodiments, the obtaining module 401 may further be configured to: when an operation event is detected, acquiring target window information; identifying the type of the target window information; when the target window information is browser window information, acquiring target element information corresponding to the browser window information through a browser plug-in and a target engine component; and when the target window information is the desktop application program window information, acquiring the target element information corresponding to the desktop application program window information through the application element identification unit.
In some embodiments, the clustering module 403 is specifically configured to: performing word segmentation processing on the plurality of behavior sequences based on a character string matching algorithm of a dictionary and/or a pre-trained machine learning model to obtain a plurality of behavior sequences after word segmentation processing; converting target window information and target element information in the behavior sequences after word segmentation into target vectors to obtain behavior sequence vector sets corresponding to the behavior sequences; and dividing each behavior sequence vector set into a plurality of behavior sequences of different categories based on a clustering algorithm.
In some embodiments, the clustering module 403 may be further configured to: acquiring the frequency or frequency of each behavior in a plurality of behavior sequences; and when the frequency or the frequency count is smaller than a first preset threshold value, deleting the behavior corresponding to the frequency or the frequency count smaller than the first preset threshold value from the behavior sequence.
In some embodiments, the clustering module 403 may be further configured to: and deleting the words of which the frequency in the behavior sequences after word segmentation is greater than a second preset threshold value.
In some embodiments, the extracting module 404 is specifically configured to: comparing each behavior event in the behavior sequence of the same category based on a preset comparison strategy, and screening out target behavior events which are consistent with each other; combining the target behavior events into a common sequence.
In some embodiments, the extraction module 404 may be further configured to: if the types of the operation events in any two behavior events are inconsistent, judging that any two behavior events are inconsistent; if the types of the operation events in any two behavior events are consistent and the event values corresponding to the operation events are inconsistent, judging that any two behavior events are inconsistent; if the target window information in any two behavior events is inconsistent, judging that any two behavior events are inconsistent; calculating the similarity between target element information in any two behavior events based on a natural language algorithm; and if the similarity is smaller than a third preset threshold value, judging that any two behavior events are inconsistent.
In some embodiments, the generating device 40 of the robot automation flow may further include: and the document generation module is used for generating a requirement definition document according to the operation events, the target window information and the target element information in the public sequence, wherein the requirement definition document is marked with specific work steps and time consumed by manual execution.
In some embodiments, the generating module 405 is specifically configured to: and defining a document according to the public sequence and the requirement, and generating an RPA flow template based on the RPA plug-in of the robot automation flow.
Each module in the apparatus shown in fig. 4 has a function of implementing each step in the method for generating a robot automation flow provided by the above method embodiment, and can achieve the corresponding technical effect, and for brevity, no further description is given here.
Based on the method for generating the robot automation process provided by the embodiment, correspondingly, the application further provides a specific implementation mode of the electronic device. Please see the examples below.
Fig. 5 shows a hardware structure diagram of an electronic device provided in an embodiment of the present application.
The electronic device may comprise a processor 501 and a memory 502 in which computer program instructions are stored.
Specifically, the processor 501 may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement the embodiments of the present Application.
In one example, memory 502 may be a Read Only Memory (ROM). In one example, the ROM can be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically Erasable PROM (EEPROM), electrically Alterable ROM (EAROM), or flash memory, or a combination of two or more of these.
The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement the method/steps in the above method embodiments, and achieve the corresponding technical effects achieved by the method/steps executed by the method embodiments, which are not described herein again for brevity.
In one example, the electronic device can also include a communication interface 503 and a bus 510. As shown in fig. 5, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 to complete communication therebetween.
The communication interface 503 is mainly used for implementing communication between modules, apparatuses, units and/or devices in the embodiments of the present application.
In addition, in combination with the method for generating the robot automation flow in the foregoing embodiments, the embodiments of the present application may be implemented by providing a computer-readable storage medium. The computer readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement the method of generating a robotic automation flow of any of the above embodiments. Examples of the computer-readable storage medium include non-transitory computer-readable storage media such as electronic circuits, semiconductor memory devices, ROMs, random access memories, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, and hard disks.
It is to be understood that the present application is not limited to the particular arrangements and instrumentality described above and shown in the attached drawings. A detailed description of known methods is omitted herein for the sake of brevity. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications, and additions or change the order between the steps after comprehending the spirit of the present application.
The functional blocks shown in the above-described structural block diagrams may be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic Circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the present application are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link. A "machine-readable medium" may include any medium that can store or transfer information. Examples of a machine-readable medium include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio Frequency (RF) links, and so forth. The code segments may be downloaded via computer networks such as the internet, intranets, etc.
It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, may be performed in an order different from the order in the embodiments, or may be performed simultaneously.
Aspects of the present application are described above with reference to flowchart illustrations and/or block diagrams of methods, apparatus, and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions/acts specified in the flowchart and/or block diagram block or blocks. Such a processor may be, but is not limited to, a general purpose processor, a special purpose processor, an application specific processor, or a field programmable logic circuit. It will also be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based computer instructions which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
As described above, only the specific embodiments of the present application are provided, and it can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system, the module and the unit described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. It should be understood that the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the present application, and these modifications or substitutions should be covered within the scope of the present application.
Claims (13)
1. A method for generating a robot automation flow, comprising:
acquiring operation data of a target user for operating the electronic equipment;
dividing the operation data into a plurality of behavior sequences according to a preset time interval;
dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, wherein the behavior sequence of each category comprises at least one behavior sequence;
extracting a public sequence of repeated work from the behavior sequences of all categories;
and generating a robot automation flow according to the public sequence.
2. The method according to claim 1, wherein the obtaining operation data of the target user operating the electronic device specifically includes:
acquiring an operation event, target window information, target element information and time of the operation event of an input device of the electronic equipment;
generating an operation log according to the operation event, the target window information, the target element information and the time of the operation event;
and obtaining the operation data according to at least one operation log.
3. The method of claim 2, wherein the target window information comprises browser window information and desktop application window information;
the acquiring the operation event, the target window information, the target element information and the time of the operation event of the input device of the electronic device specifically includes:
when the operation event is detected, acquiring the target window information;
identifying a type of the target window information;
when the target window information is the browser window information, acquiring the target element information corresponding to the browser window information through a browser plug-in and a target engine component;
and when the target window information is the desktop application program window information, acquiring the target element information corresponding to the desktop application program window information through an application element identification unit.
4. The method according to claim 1, wherein the dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm specifically comprises:
performing word segmentation processing on the plurality of behavior sequences based on a character string matching algorithm of a dictionary and/or a pre-trained machine learning model to obtain the plurality of behavior sequences after word segmentation processing;
converting target window information and target element information in the behavior sequences after word segmentation into target vectors to obtain behavior sequence vector sets corresponding to the behavior sequences respectively;
and dividing each behavior sequence vector set into a plurality of behavior sequences of different categories based on a clustering algorithm.
5. The method according to claim 4, wherein before the performing word segmentation processing on the plurality of behavior sequences, the dividing the plurality of behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm further includes:
acquiring the frequency or frequency of each behavior in the plurality of behavior sequences;
and when the frequency or the frequency count is smaller than a first preset threshold value, deleting the behavior corresponding to the frequency or the frequency count smaller than the first preset threshold value from the behavior sequence.
6. The method according to claim 4, wherein after the performing word segmentation processing on the behavior sequences, the dividing the behavior sequences into behavior sequences of different categories based on a clustering algorithm further includes:
deleting the words of which the frequency in the behavior sequences after word segmentation is greater than a second preset threshold value.
7. The method of claim 1, wherein the sequence of behaviors comprises a plurality of behavioral events;
the extracting of the common sequence of the repetitive work from the behavior sequences of each category specifically includes:
comparing all behavior events in the behavior sequences of the same category based on a preset comparison strategy, and screening out target behavior events which are consistent with each other;
combining the target behavior events into the common sequence.
8. The method according to claim 7, wherein the behavior event includes an operation event of an input device of the electronic apparatus, target window information, target element information, and a time of the operation event;
the method includes the steps of comparing all behaviors in a behavior sequence of the same category based on a preset comparison strategy, and screening out target behaviors which are consistent with each other, and specifically includes the following steps:
if the types of the operation events in any two behavior events are inconsistent, judging that the any two behavior events are inconsistent;
if the types of the operation events in any two behavior events are consistent, and the event values corresponding to the operation events are inconsistent, judging that the any two behavior events are inconsistent;
if the target window information in any two behavior events is inconsistent, judging that the any two behavior events are inconsistent;
calculating the similarity between the target element information in any two behavior events based on a natural language algorithm;
and if the similarity is smaller than a third preset threshold value, judging that the any two behavior events are inconsistent.
9. The method of claim 1, further comprising, after extracting the common sequence of repetitive work from the sequence of behaviors of each category:
and generating a requirement definition document according to the operation events, the target window information and the target element information in the public sequence, wherein the requirement definition document is marked with specific work steps and time consumed for manual execution.
10. The method according to claim 9, wherein generating a robotic automation process from the common sequence comprises:
and defining a document according to the public sequence and the requirement, and generating an RPA flow template based on a robot automation flow RPA plug-in.
11. An apparatus for generating a robot automation flow, comprising:
the acquisition module is used for acquiring operation data of a target user for operating the electronic equipment;
the dividing module is used for dividing the operation data into a plurality of behavior sequences according to a preset time interval;
the clustering module is used for dividing the behavior sequences into a plurality of behavior sequences of different categories based on a clustering algorithm, and the behavior sequence of each category comprises at least one behavior sequence;
the extraction module is used for extracting a public sequence of repeated work from the behavior sequences of all categories;
and the generating module is used for generating the robot automation flow according to the public sequence.
12. An electronic device, characterized in that the electronic device comprises: processor, memory and computer program stored on the memory and executable on the processor, which computer program, when executed by the processor, carries out the steps of the method of generating a robotic automation flow as claimed in any one of claims 1 to 10.
13. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, which computer program, when being executed by a processor, carries out the steps of the method for generating a robotic automation flow as claimed in any one of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211630249.1A CN115953123A (en) | 2022-12-19 | 2022-12-19 | Method, device and equipment for generating robot automation flow and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211630249.1A CN115953123A (en) | 2022-12-19 | 2022-12-19 | Method, device and equipment for generating robot automation flow and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115953123A true CN115953123A (en) | 2023-04-11 |
Family
ID=87285424
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211630249.1A Pending CN115953123A (en) | 2022-12-19 | 2022-12-19 | Method, device and equipment for generating robot automation flow and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115953123A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117406972A (en) * | 2023-12-14 | 2024-01-16 | 安徽思高智能科技有限公司 | RPA high-value flow instance discovery method and system based on fitness analysis |
| CN117592622A (en) * | 2024-01-19 | 2024-02-23 | 西南科技大学 | Robot flow automation-oriented behavior sequence prediction method and system |
| CN117592622B (en) * | 2024-01-19 | 2024-04-30 | 西南科技大学 | Robot flow automation-oriented behavior sequence prediction method and system |
-
2022
- 2022-12-19 CN CN202211630249.1A patent/CN115953123A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117406972A (en) * | 2023-12-14 | 2024-01-16 | 安徽思高智能科技有限公司 | RPA high-value flow instance discovery method and system based on fitness analysis |
| CN117406972B (en) * | 2023-12-14 | 2024-02-13 | 安徽思高智能科技有限公司 | RPA high-value flow instance discovery method and system based on fitness analysis |
| CN117592622A (en) * | 2024-01-19 | 2024-02-23 | 西南科技大学 | Robot flow automation-oriented behavior sequence prediction method and system |
| CN117592622B (en) * | 2024-01-19 | 2024-04-30 | 西南科技大学 | Robot flow automation-oriented behavior sequence prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113011533B (en) | Text classification method, apparatus, computer device and storage medium | |
| CN110851598B (en) | Text classification method and device, terminal equipment and storage medium | |
| CN109189767B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN111444723A (en) | Information extraction model training method and device, computer equipment and storage medium | |
| CN111651601B (en) | Training method and classification method for fault classification model of power information system | |
| CN110175334B (en) | Text knowledge extraction system and method based on custom knowledge slot structure | |
| CN108027814B (en) | Stop word recognition method and device | |
| CN107577702B (en) | Method for distinguishing traffic information in social media | |
| CN111177367B (en) | Case classification method, classification model training method and related products | |
| CN111274817A (en) | Intelligent software cost measurement method based on natural language processing technology | |
| CN110008474B (en) | Key phrase determining method, device, equipment and storage medium | |
| CN109783636B (en) | Automobile comment theme extraction method based on classifier chain | |
| CN110580308A (en) | information auditing method and device, electronic equipment and storage medium | |
| CN113221918B (en) | Target detection method, training method and device of target detection model | |
| CN111259951A (en) | Case detection method and device, electronic equipment and readable storage medium | |
| CN111177375A (en) | Electronic document classification method and device | |
| CN113986950A (en) | SQL statement processing method, device, equipment and storage medium | |
| CN115859980A (en) | Semi-supervised named entity identification method, system and electronic equipment | |
| CN115146062A (en) | Intelligent event analysis method and system fusing expert recommendation and text clustering | |
| CN116150201A (en) | Sensitive data identification method, device, equipment and computer storage medium | |
| CN113486178B (en) | Text recognition model training method, text recognition method, device and medium | |
| CN115953123A (en) | Method, device and equipment for generating robot automation flow and storage medium | |
| CN114547301A (en) | Document processing method, document processing device, recognition model training equipment and storage medium | |
| CN112579781B (en) | Text classification method, device, electronic equipment and medium | |
| CN113935314A (en) | Abstract extraction method, device, terminal equipment and medium based on heteromorphic graph network |
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 |