CN114547255A

CN114547255A - Conversation prompting method, device, storage medium and equipment based on artificial intelligence

Info

Publication number: CN114547255A
Application number: CN202210238280.4A
Authority: CN
Inventors: 翟永青
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2022-03-10
Filing date: 2022-03-10
Publication date: 2022-05-27

Abstract

The invention belongs to the technical field of artificial intelligence, and particularly relates to a session prompting method and device based on artificial intelligence, a computer readable storage medium and terminal equipment. The method comprises the following steps: extracting a historical conversation text set between an agent and a user from a preset database; performing vocabulary frequent sequence mining on each section of historical session text in the historical session text set to obtain a vocabulary frequent sequence set corresponding to the historical session text set; acquiring target keywords of current seat personnel and a user in a real-time conversation process; searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set; and carrying out conversation prompt on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset. According to the invention, the historical conversation between the seat personnel and the user is mined to prompt the conversation in real time without making a script in advance, and the support can be flexibly provided for the seat personnel in various scenes.

Description

Conversation prompting method, device, storage medium and equipment based on artificial intelligence

Technical Field

The invention belongs to the technical field of artificial intelligence, and particularly relates to a session prompting method and device based on artificial intelligence, a computer readable storage medium and terminal equipment.

Background

Modern seating scenes have increasingly shifted from offline to online, including telephone seating, online customer service, and the like; these agents have one thing in common: the conversation with the user by using the natural language requires a certain expertise, but because the conversation with the user is not face-to-face, the seat personnel can look up the data to a certain extent and delay the reply.

In the prior art, the seat personnel generally carry out conversation with the user according to a preset script, but the flexibility of the method is poor, when some content which is not covered in the script appears in the conversation process, the seat personnel cannot be supported by the script, and can only carry out conversation with the user by virtue of personal experience of the seat personnel, so that high-quality service experience is difficult to provide for the user.

Disclosure of Invention

In view of this, embodiments of the present invention provide a session prompting method and apparatus based on artificial intelligence, a computer-readable storage medium, and a terminal device, so as to solve the problem that an agent has poor flexibility when performing a session with a user according to a preset script.

A first aspect of an embodiment of the present invention provides a session prompting method based on artificial intelligence, which may include:

extracting a historical conversation text set between an agent person and a user from a preset database, wherein the historical conversation text set comprises various historical conversation texts in a preset statistical time period;

performing vocabulary frequent sequence mining on each section of historical session text in the historical session text set to obtain a vocabulary frequent sequence set corresponding to the historical session text set;

acquiring target keywords of current seat personnel and a user in a real-time conversation process;

searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set;

and carrying out conversation prompt on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset.

In a specific implementation manner of the first aspect, the database may include more than two sub-libraries, where each sub-library corresponds to one professional field in a preset professional field set;

the extracting of the historical conversation text set between the agent person and the user from the preset database may include:

searching a target sub-library corresponding to a first professional field from the database, wherein the first professional field is any one of the professional fields in the professional field set;

extracting a historical conversation text subset between the agent personnel and the user in the first professional field from the target sub-library;

the mining of frequent sequences of words for each section of historical session text in the historical session text set comprises:

and performing vocabulary frequent sequence mining on each section of historical conversation text in the historical conversation text subset of the first professional field to obtain a vocabulary frequent sequence set of the first professional field.

In a specific implementation manner of the first aspect, the searching, in the set of frequent vocabulary sequences, a subset of frequent vocabulary sequences corresponding to the target keyword may include:

acquiring the identity of the current seat personnel;

inquiring a second professional field in a preset agent professional field distribution list according to the identity, wherein the second professional field is a professional field corresponding to the current agent;

and searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set of the second professional field.

In a specific implementation manner of the first aspect, the obtaining target keywords of the current agent staff and the user in the real-time conversation process may include:

acquiring input words of the current seat personnel through a preset human-computer interaction interface, and determining the input words as the target keywords;

or

Carrying out automatic voice recognition on voice of the current seat personnel and the user in the real-time conversation process to obtain a real-time conversation text;

performing word segmentation processing on the real-time conversation text to obtain a word segmentation sequence;

determining each word in the word segmentation sequence as the target keyword.

In a specific implementation manner of the first aspect, the searching, in the frequent vocabulary sequence set, a frequent vocabulary sequence subset corresponding to the target keyword may include:

selecting an unselected vocabulary frequent sequence from the vocabulary frequent sequence set as a candidate vocabulary frequent sequence;

searching the target keyword in the candidate vocabulary frequent sequence;

if the target keyword is found in the candidate frequent vocabulary sequence, adding the candidate frequent vocabulary sequence into the frequent vocabulary sequence subset;

and if the target keyword cannot be searched in the candidate frequent vocabulary sequence, returning to execute the step of selecting a frequently-selected vocabulary sequence from the frequent vocabulary sequence set as the candidate frequent vocabulary sequence until the frequently-selected vocabulary sequence does not exist in the frequent vocabulary sequence set.

In a specific implementation manner of the first aspect, the performing a conversation prompt on the current seat person according to each frequent vocabulary sequence in the frequent vocabulary sequence subset may include:

respectively inquiring the priority index of each vocabulary frequent sequence in the vocabulary frequent sequence subset in a preset priority index list;

sequencing the vocabulary frequent sequences in the vocabulary frequent sequence subset according to the sequence of the priority indexes from large to small to obtain a sequenced vocabulary frequent sequence subset;

and carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the sorted vocabulary frequent sequence subset.

In a specific implementation manner of the first aspect, the setting process of the priority index list may include:

respectively inquiring the use times of each vocabulary frequent sequence in a preset vocabulary frequent sequence use record;

respectively determining the priority index of each vocabulary frequent sequence according to the using times of each vocabulary frequent sequence;

and constructing the priority index list according to the priority indexes of the frequent sequences of the vocabularies.

A second aspect of the embodiments of the present invention provides a session prompt apparatus based on artificial intelligence, which may include:

the system comprises a session text extraction module, a database processing module and a database processing module, wherein the session text extraction module is used for extracting a historical session text set between an agent and a user from a preset database, and the historical session text set comprises various historical session texts in a preset statistical time period;

the vocabulary frequent sequence mining module is used for mining the vocabulary frequent sequence of each section of historical session text in the historical session text set to obtain a vocabulary frequent sequence set corresponding to the historical session text set;

the target keyword acquisition module is used for acquiring target keywords of current seat personnel and a user in a real-time conversation process;

the sequence subset searching module is used for searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set;

and the conversation prompting module is used for carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset.

In a specific implementation manner of the second aspect, the database may include more than two sub-libraries, where each sub-library corresponds to one professional field in a preset professional field set;

the session text extraction module may specifically include:

a target sub-library searching unit, configured to search a target sub-library corresponding to a first professional field from the database, where the first professional field is any one of the professional fields in the professional field set;

the conversation text subset extraction unit is used for extracting a historical conversation text subset between the agent personnel and the user in the first professional field from the target sub-library;

the frequent sequence of vocabulary mining module may include:

and the first mining unit is used for mining frequent vocabulary sequences of all sections of historical conversation texts in the historical conversation text subset of the first professional field to obtain a frequent vocabulary sequence set of the first professional field.

In a specific implementation manner of the second aspect, the sequence subset searching module may specifically include:

the identity identification obtaining unit is used for obtaining the identity identification of the current seat personnel;

the professional field query unit is used for querying a second professional field in a preset agent professional field distribution list according to the identity, wherein the second professional field is a professional field corresponding to the current agent;

and the vocabulary frequent sequence subset searching unit is used for searching the vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set in the second professional field.

In a specific implementation manner of the second aspect, the target keyword obtaining module may be specifically configured to: acquiring input words of the current seat personnel through a preset human-computer interaction interface, and determining the input words as the target keywords; or, carrying out automatic voice recognition on the voice of the current seat personnel and the user in the real-time conversation process to obtain a real-time conversation text; performing word segmentation processing on the real-time conversation text to obtain a word segmentation sequence; determining each word in the word segmentation sequence as the target keyword.

In a specific implementation manner of the second aspect, the sequence subset searching module may be specifically configured to: selecting an unselected vocabulary frequent sequence from the vocabulary frequent sequence set as a candidate vocabulary frequent sequence; searching the target keyword in the candidate vocabulary frequent sequence; if the target keyword is found in the candidate frequent vocabulary sequence, adding the candidate frequent vocabulary sequence into the frequent vocabulary sequence subset; and if the target keyword cannot be searched in the candidate frequent vocabulary sequence, returning to execute the step of selecting a frequently-selected vocabulary sequence from the frequent vocabulary sequence set as the candidate frequent vocabulary sequence until the frequently-selected vocabulary sequence does not exist in the frequent vocabulary sequence set.

In a specific implementation manner of the second aspect, the session prompt module may be specifically configured to: respectively inquiring the priority index of each vocabulary frequent sequence in the vocabulary frequent sequence subset in a preset priority index list; sequencing the vocabulary frequent sequences in the vocabulary frequent sequence subset according to the sequence of the priority indexes from large to small to obtain a sequenced vocabulary frequent sequence subset; and carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the sorted vocabulary frequent sequence subset.

In a specific implementation manner of the second aspect, the session prompting apparatus may further include:

the priority index list setting module is used for respectively inquiring the use times of each vocabulary frequent sequence in a preset vocabulary frequent sequence use record; respectively determining the priority index of each vocabulary frequent sequence according to the using times of each vocabulary frequent sequence; and constructing the priority index list according to the priority indexes of the frequent sequences of the vocabularies.

A third aspect of the embodiments of the present invention provides a computer-readable storage medium, where a computer program is stored, and the computer program, when executed by a processor, implements the steps of any one of the above-mentioned session prompting methods.

A fourth aspect of the embodiments of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements any of the steps of the session prompting method when executing the computer program.

A fifth aspect of embodiments of the present invention provides a computer program product, which, when running on a terminal device, causes the terminal device to perform any of the steps of the session prompt method described above.

Compared with the prior art, the embodiment of the invention has the following beneficial effects: extracting a historical conversation text set between an agent and a user from a preset database, wherein the historical conversation text set comprises all sections of historical conversation texts in a preset statistical time period; performing vocabulary frequent sequence mining on each section of historical session text in the historical session text set to obtain a vocabulary frequent sequence set corresponding to the historical session text set; acquiring target keywords of current seat personnel and a user in a real-time conversation process; searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set; and carrying out conversation prompt on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset. According to the embodiment of the invention, the historical conversation between the seat personnel and the user is mined to prompt the conversation in real time without making a script in advance, and the support can be flexibly provided for the seat personnel in various scenes.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

Fig. 1 is a flowchart of an embodiment of a session prompt method according to an embodiment of the present invention;

FIG. 2 is a schematic flow diagram of a conversation prompt for a current agent based on each frequent sequence of vocabulary in the frequent sequence subset of vocabularies;

FIG. 3 is a block diagram of an embodiment of a session prompt apparatus according to an embodiment of the present invention;

fig. 4 is a schematic block diagram of a terminal device in an embodiment of the present invention.

Detailed Description

In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the invention can acquire and process related data based on an artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.

The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.

The execution subject of the embodiment of the present invention may be a terminal device based on artificial intelligence, and is used to execute the session prompt method in the embodiment of the present invention. Referring to fig. 1, an embodiment of a session prompting method based on artificial intelligence in an embodiment of the present invention may include:

step S101, extracting a historical conversation text set between the seat personnel and the user from a preset database.

The historical conversation text set comprises each historical conversation text in a preset statistical time period, and the statistical time period can be set according to actual conditions, for example, a time period of one week, one month or one quarter from the current time can be used as the statistical time period.

In the statistical time period, the terminal device may continuously collect session texts between each agent and each user in the session process of each time, and store the session texts as historical session texts in the database for subsequent analysis.

Preferably, each piece of historical conversation text stored in the database can be continuously updated and overlaid, that is, as time goes on, when new conversation text needs to be stored and the storage space is insufficient, the conversation text with the earliest storage time can be overlaid, so that the conversation text which is closest to the current moment in the time dimension is ensured to be stored in the database. It is easy to understand that, because products and services of enterprises are generally in the process of continuous updating, the conversation content between the agent personnel and the user is changed correspondingly, and therefore, the closer the conversation text is to the current moment, the higher the reference value is.

When the historical conversation texts need to be analyzed, the terminal device can extract each section of historical conversation texts from the database, and a set formed by the historical conversation texts is the historical conversation text set.

And step S102, performing vocabulary frequent sequence mining on each section of historical session text in the historical session text set to obtain a vocabulary frequent sequence set corresponding to the historical session text set.

And for each section of historical conversation text, respectively carrying out character removal processing, word segmentation processing, stop word removal processing and the like on the historical conversation text, and converting each section of historical conversation text into a sequence consisting of a plurality of word segments.

The character removing processing refers to removing the contents of numbers, punctuations, letters, symbols and the like in the text and only keeping the Chinese characters in the text.

The word segmentation processing means that the text is segmented into a single word, and each segmented word is marked as a word segmentation. In the embodiment of the invention, the text can be segmented according to the universal dictionary, so that the segmented words are all normal words, and if the words are not in the dictionary, the single words are segmented. When words can be formed in the front and back directions, for example, "ABC" can be divided according to the size of the statistical word frequency, if the word frequency of "AB" is high, "AB/C" is divided, and if the word frequency of "BC" is high, "A/BC" is divided.

Stop words refer to words that are typically filtered out before processing natural language, which are the most common words in virtually any language, such as articles, prepositions, pronouns, conjunctions, etc., without adding too much information to the text. Stop words are abundant in any human language. By deleting these words, the underlying information can be deleted from the text to focus more on important information. Moreover, removing stop words can greatly reduce the data volume, thereby improving the efficiency of text processing. In the embodiment of the invention, the stop word removing processing can be carried out on the text according to the preset stop word list, and stop words appearing after the text is segmented are deleted from the segmentation sequence.

After the processing, each section of the historical conversation text is converted into a word segmentation sequence, a set formed by the sequences is marked as a sequence data set S, and then the sequence data set S is processed by using a preset frequent vocabulary sequence mining algorithm to obtain a frequent vocabulary sequence set corresponding to the historical conversation text set.

In the embodiment of the invention, any frequent vocabulary sequence mining algorithm can be selected according to actual conditions, a Prefix span algorithm is taken as an example for explanation, the Prefix span is a depth-first search algorithm using a database projection technology, and has the capability of processing a large sequence database. And then, a divide-and-conquer strategy is adopted, more and smaller projection databases are continuously generated, and the sequence mode mining is carried out on each projection database. The input of the method is a sequence data set S and a support degree threshold value min _ sup, the output of the method is a frequent vocabulary sequence set meeting the support degree requirement, and the variables are explained as follows:

alpha: a prefix;

l: prefix alpha length;

s | alpha: if alpha is not null, S | alpha is the data set S projected by alpha, otherwise S | alpha is the sequence data set S.

The detailed process of the Prefix span algorithm includes:

(1) finding out all alpha (prefix) with the length of 1 and corresponding S | alpha (projection database);

(2) counting alpha (prefix) with the length of 1, deleting items corresponding to prefixes with the support degrees lower than a threshold value min _ sup from a data set S, and simultaneously obtaining all sequences (b1) with 1 item frequently, wherein i is 1;

(3) performing recursive mining on alpha (prefix) with the length of i meeting the requirement of the support degree min _ sup:

a) find S | A' (projection database) corresponding to A (prefix). If the projection database is empty (empty set), then recursively return;

b) and counting the support degree counts of all items in the corresponding projection database. If the support counts of all the entries are below the threshold α, then the recursion returns;

c) combining each single item meeting the support degree min _ sup count with the current prefix to obtain a plurality of new prefixes;

d) and (4) making i equal to i +1, wherein the prefixes are the prefixes after the single items are merged, and respectively executing the step (3) in a recursive manner.

It should be noted that the above PrefixSpan algorithm is only an example, and is not limited, and in practical applications, other vocabulary frequent sequence mining algorithms may be selected according to practical situations to perform vocabulary frequent sequence mining, so as to obtain a vocabulary frequent sequence set corresponding to the historical conversation text set.

And S103, acquiring target keywords of the current seat personnel and the user in the real-time conversation process.

The current seat person may be any one of the seat persons. In a specific implementation of the embodiment of the present invention, in a process of a real-time conversation between the current agent and the user, if some content not covered in the script is encountered and cannot be supported by the script, a related word (denoted as an input word) may be input through a preset human-computer interaction interface, and the terminal device may obtain the input word through the human-computer interaction interface and determine the input word as a target keyword. For example, if the user refers to the product a in the session and the script does not have the content related to the product a, the current agent person may input the name of the product a as an input word into the human-computer interaction interface and transmit the input word to the terminal device.

In another specific implementation of the embodiment of the present invention, the terminal device may further automatically obtain a target keyword in a real-time session.

Firstly, Automatic Speech Recognition (ASR) is performed on Speech of a current agent and a user in a real-time conversation process to obtain a corresponding real-time conversation text. The automatic speech recognition system used in the embodiment of the invention can comprise four parts, namely feature extraction, an acoustic model, a language model, a dictionary and decoding, and can also carry out audio data preprocessing work such as filtering, framing and the like on speech in order to effectively extract features, and appropriately extract an audio signal to be analyzed from an original signal; the feature extraction work converts the voice from a time domain to a frequency domain, and provides a proper feature vector for the acoustic model; calculating the score of each feature vector on the acoustic features according to the acoustic characteristics in the acoustic model; the language model calculates the probability of the possible phrase sequence corresponding to the voice according to the theory related to linguistics; and finally, decoding the phrase sequence according to the existing dictionary to obtain the final possible text.

After the real-time conversation text is obtained, the terminal device may perform character removal processing, word segmentation processing, stop word removal processing, and the like on the real-time conversation text, respectively, to obtain a processed word segmentation sequence, and determine each word in the word segmentation sequence as the target keyword. In this case, as the real-time conversation process between the current agent person and the user, each word in the real-time conversation process is sequentially executed in step 104, so that the conversation prompt for the current agent person is dynamically realized.

And step S104, searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set.

In an initial state, the vocabulary frequent sequence subset is empty. Firstly, selecting an unselected vocabulary frequent sequence from the vocabulary frequent sequence set as a candidate vocabulary frequent sequence. In the embodiment of the invention, the candidate vocabulary frequent sequence can be randomly selected or sequentially selected according to the sequence. Then, the target keywords are searched in the candidate vocabulary frequent sequence. If the target keyword is found in the candidate frequent vocabulary sequence, adding the candidate frequent vocabulary sequence into the frequent vocabulary sequence subset; and if the target keyword cannot be searched in the candidate frequent vocabulary sequence, returning to execute the step of selecting a frequently-selected vocabulary sequence from the frequent vocabulary sequence set as the candidate frequent vocabulary sequence until the frequently-selected vocabulary sequence does not exist in the frequent vocabulary sequence set.

And S105, carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset.

After the vocabulary frequent sequence subset is obtained, the terminal device can display each vocabulary frequent sequence in the vocabulary frequent sequence subset on a preset human-computer interaction interface so as to prompt the current seat personnel for conversation.

In a specific implementation manner of the embodiment of the present invention, step S105 may specifically include a process as shown in fig. 2:

step S1051, respectively inquiring the priority index of each vocabulary frequent sequence in the vocabulary frequent sequence subset in a preset priority index list.

In the embodiment of the present invention, the priority index list may be set in advance. Specifically, the using times of each vocabulary frequent sequence are respectively inquired in a preset vocabulary frequent sequence using record, and then the priority index of each vocabulary frequent sequence is respectively determined according to the using times of each vocabulary frequent sequence, wherein the priority index of any one vocabulary frequent sequence is positively correlated with the using times of the vocabulary frequent sequence, the priority index is larger when the using times of one vocabulary frequent sequence are more, and the priority index is smaller when the using times of one vocabulary frequent sequence are less. Finally, the priority index list may be constructed from the priority indexes of the frequent sequences of words.

And step 1052, sequencing the vocabulary frequent sequences in the vocabulary frequent sequence subset according to the sequence of the priority indexes from large to small to obtain a sequenced vocabulary frequent sequence subset.

Through the sequencing, the vocabulary frequent sequence with more use times is arranged at the position which is the front, and the vocabulary frequent sequence with less use times is arranged at the position which is the rear, so that the current seat personnel can conveniently obtain the most useful conversation promotion in time.

And S1053, carrying out conversation prompt on the current seat personnel according to each vocabulary frequent sequence in the sorted vocabulary frequent sequence subset.

Because human beings have the ability of automatic word connection sentence making, only the corresponding key word prompt is needed for the seat personnel, the seat can organize words into a conversation by self, and the conversation does not need to be completely carried out according to scripts. Therefore, the embodiment of the invention excavates the keyword information in past historical conversations to give other high-frequency vocabulary information which is appeared when a topic is talked, the high-frequency vocabulary information is often the keyword which facilitates the topic to be smoothly carried out, and the keyword can enable the conversation of the seat personnel.

Further, in consideration of the fact that the agent personnel may have more and more detailed professional field division and there may be a great difference between different professional fields (such as insurance, stock, fund, etc.), in a specific implementation of the embodiment of the present invention, mutually independent conversation prompting systems may be provided for the agent personnel in different professional fields respectively.

In this case, the database may be divided into two or more sub-libraries, where each sub-library corresponds to one professional field in a preset professional field set, all the agent personnel may also be divided into each professional field, and each agent personnel only handles services corresponding to its professional field.

During the statistical time period, the terminal device may continuously collect session texts between each agent and each user during the previous session, store the session texts as historical session texts into the sub-library corresponding to the professional field of the terminal device, for example, the session texts between the agent and the user in the insurance professional field may be stored into the sub-library corresponding to the insurance professional field, the session texts between the agent and the user in the stock professional field may be stored into the sub-library corresponding to the stock professional field, the session texts between the agent and the user in the fund professional field may be stored into the sub-library corresponding to the fund professional field, and so on.

When the historical conversation text needs to be analyzed, the terminal equipment can analyze each professional field respectively. Specifically, taking any one of the professional fields (denoted as a first professional field) in the set of professional fields as an example, when extracting the historical conversation text, the terminal device may search a target sub-library corresponding to the first professional field from the database, and extract a historical conversation text subset between the agent person and the user in the first professional field from the target sub-library. When performing frequent vocabulary sequence mining, the terminal device may perform frequent vocabulary sequence mining on each section of the historical conversation text in the subset of the historical conversation texts in the first professional field, so as to obtain the frequent vocabulary sequence set in the first professional field. For example, the terminal device may extract each section of historical conversation text from a sub-library corresponding to the fund specialty field, and a set formed by these historical conversation texts is a set of historical conversation texts corresponding to the fund specialty field. And performing frequent vocabulary sequence mining on the vocabulary sets to obtain frequent vocabulary sequence sets corresponding to the fund professional fields, and by analogy, obtaining frequent vocabulary sequence sets corresponding to the professional fields respectively.

When searching for frequent vocabulary sequence subsets, the terminal device may first obtain an identity of the current agent, query a second professional field in a preset allocation list of professional fields of the agent according to the identity, where the second professional field is a professional field corresponding to the current agent, and search for the frequent vocabulary sequence subset corresponding to the target keyword from the frequent vocabulary sequence set in the second professional field. And finally, carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset. Through the mode, mutual interference among different professional fields is avoided, and conversation prompt is more accurate and effective.

In summary, in the embodiments of the present invention, a historical conversation text set between an agent and a user is extracted from a preset database, where the historical conversation text set includes various segments of historical conversation texts within a preset statistical time period; performing vocabulary frequent sequence mining on each section of historical conversation text in the historical conversation text set to obtain a vocabulary frequent sequence set corresponding to the historical conversation text set; acquiring target keywords of current seat personnel and a user in a real-time conversation process; searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set; and carrying out conversation prompt on the current seat personnel according to each vocabulary frequent sequence in the vocabulary frequent sequence subset. According to the embodiment of the invention, the historical conversation between the seat personnel and the user is mined to prompt the conversation in real time without making a script in advance, and the support can be flexibly provided for the seat personnel in various scenes.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

Fig. 3 is a structural diagram of an embodiment of an artificial intelligence based session prompt apparatus according to an embodiment of the present invention, which corresponds to the artificial intelligence based session prompt method according to the foregoing embodiment.

In this embodiment, a session prompting device based on artificial intelligence may include:

a session text extraction module 301, configured to extract a historical session text set between an agent and a user from a preset database, where the historical session text set includes various segments of historical session texts within a preset statistical time period;

a vocabulary frequent sequence mining module 302, configured to perform vocabulary frequent sequence mining on each section of historical session text in the historical session text set, so as to obtain a vocabulary frequent sequence set corresponding to the historical session text set;

a target keyword obtaining module 303, configured to obtain a target keyword of a current agent and a user in a real-time session;

a sequence subset searching module 304, configured to search, in the vocabulary frequent sequence set, a vocabulary frequent sequence subset corresponding to the target keyword;

and the conversation prompting module 305 is configured to perform conversation prompting on the current seat person according to each frequent vocabulary sequence in the frequent vocabulary sequence subset.

In a specific implementation manner of the embodiment of the present invention, the database may include more than two sub-libraries, where each sub-library corresponds to one professional field in a preset professional field set;

the session text extraction module may specifically include:

the frequent sequence of vocabulary mining module may include:

In a specific implementation manner of the embodiment of the present invention, the sequence subset search module may specifically include:

In a specific implementation manner of the embodiment of the present invention, the target keyword obtaining module may be specifically configured to: acquiring input words of the current seat personnel through a preset human-computer interaction interface, and determining the input words as the target keywords; or, carrying out automatic voice recognition on the voice of the current seat personnel and the user in the real-time conversation process to obtain a real-time conversation text; performing word segmentation processing on the real-time conversation text to obtain a word segmentation sequence; determining each word in the word segmentation sequence as the target keyword.

In a specific implementation manner of the embodiment of the present invention, the sequence subset search module may be specifically configured to: selecting an unselected vocabulary frequent sequence from the vocabulary frequent sequence set as a candidate vocabulary frequent sequence; searching the target keyword in the candidate vocabulary frequent sequence; if the target keyword is found in the candidate frequent vocabulary sequence, adding the candidate frequent vocabulary sequence into the frequent vocabulary sequence subset; and if the target keyword cannot be searched in the candidate frequent vocabulary sequence, returning to execute the step of selecting a frequently-selected vocabulary sequence from the frequent vocabulary sequence set as the candidate frequent vocabulary sequence until the frequently-selected vocabulary sequence does not exist in the frequent vocabulary sequence set.

In a specific implementation manner of the embodiment of the present invention, the session prompt module may be specifically configured to: respectively inquiring the priority index of each vocabulary frequent sequence in the vocabulary frequent sequence subset in a preset priority index list; sequencing the vocabulary frequent sequences in the vocabulary frequent sequence subset according to the sequence of the priority indexes from large to small to obtain a sequenced vocabulary frequent sequence subset; and carrying out conversation prompting on the current seat personnel according to each vocabulary frequent sequence in the sorted vocabulary frequent sequence subset.

In a specific implementation manner of the embodiment of the present invention, the session prompt apparatus may further include:

It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described apparatuses, modules and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Fig. 4 shows a schematic block diagram of a terminal device according to an embodiment of the present invention, and for convenience of description, only the parts related to the embodiment of the present invention are shown.

In this embodiment, the terminal device 4 may include: a processor 40, a memory 41, and computer readable instructions 42 stored in the memory 41 and executable on the processor 40, such as computer readable instructions to perform the above-described session alert method. The processor 40, when executing the computer readable instructions 42, implements the steps in the various embodiments of the session prompting method described above, such as the steps S101 to S105 shown in fig. 1. Alternatively, the processor 40, when executing the computer readable instructions 42, implements the functions of the modules/units in the above device embodiments, such as the functions of the modules 301 to 305 shown in fig. 3.

Illustratively, the computer readable instructions 42 may be partitioned into one or more modules/units that are stored in the memory 41 and executed by the processor 40 to implement the present invention. The one or more modules/units may be a series of computer-readable instruction segments capable of performing specific functions, which are used for describing the execution process of the computer-readable instructions 42 in the terminal device 4.

It will be understood by those skilled in the art that fig. 4 is only an example of the terminal device 4, and does not constitute a limitation to the terminal device 4, and may include more or less components than those shown, or combine some components, or different components, for example, the terminal device 4 may further include an input-output device, a network access device, a bus, etc.

The Processor 40 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 41 may be an internal storage unit of the terminal device 4, such as a hard disk or a memory of the terminal device 4. The memory 41 may also be an external storage device of the terminal device 4, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the terminal device 4. Further, the memory 41 may also include both an internal storage unit and an external storage device of the terminal device 4. The memory 41 is used for storing the computer readable instructions and other instructions and data required by the terminal device 4. The memory 41 may also be used to temporarily store data that has been output or is to be output.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other ways. For example, the above-described embodiments of the apparatus/terminal device are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow in the method of the embodiments described above can be realized by a computer program, which can be stored in a computer-readable storage medium and can realize the steps of the embodiments of the methods described above when the computer program is executed by a processor. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable storage medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable storage media that does not include electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A conversation prompting method based on artificial intelligence is characterized by comprising the following steps:

extracting a historical conversation text set between an agent and a user from a preset database, wherein the historical conversation text set comprises all sections of historical conversation texts in a preset statistical time period;

2. The conversation prompting method according to claim 1, wherein the database comprises more than two sub-libraries, wherein each sub-library corresponds to one professional field in a preset professional field set;

the extracting of the historical conversation text set between the agent personnel and the user from the preset database includes:

3. The conversation prompting method according to claim 2, wherein the searching for the vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set comprises:

acquiring the identity of the current seat personnel;

and searching a vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set in the second professional field.

4. The conversation prompting method according to claim 1, wherein the obtaining of the target keyword of the current agent person and the user in the real-time conversation process comprises:

or

determining each word in the word segmentation sequence as the target keyword.

5. The conversation prompting method according to claim 1, wherein the searching for the vocabulary frequent sequence subset corresponding to the target keyword in the vocabulary frequent sequence set comprises:

searching the target keyword in the candidate vocabulary frequent sequence;

6. The conversation prompting method according to any one of claims 1 to 5, wherein the conversation prompting of the current agent person according to each vocabulary frequent sequence in the vocabulary frequent sequence subset comprises:

7. The session prompt method according to claim 6, wherein the setting process of the priority index list comprises:

8. A conversation prompting device based on artificial intelligence is characterized by comprising:

9. A computer readable storage medium storing computer readable instructions, which when executed by a processor implement the steps of the session alerting method of any one of claims 1-7.

10. A terminal device comprising a memory, a processor and computer readable instructions stored in the memory and executable on the processor, wherein the processor when executing the computer readable instructions implements the steps of the session alerting method of any one of claims 1-7.