A new customer identification method and device based on finite state machine
Through the finite state machine segmented different customers in offline sales scenarios, the problem of the existing technology being difficult to distinguish customers when short recording and voiceprint information is incomplete, and efficient and accurate identification of new customers is achieved, suitable for cost-sensitive business scenarios.
Patent Information
- Application Number
- CN202110913469.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-15
- Publication Date
- 2024-12-13
- Estimated Expiration
- 2041-10-15
AI Technical Summary
In offline sales scenarios, it is difficult for the existing technology to effectively distinguish different customers, especially when the recording time is short and the voiceprint information is incomplete, and the existing machine learning model depends on feature engineering and has low accuracy, especially in cosmetics. There are problems of waiting time and strong interference in the shopping guide scenario.
A new customer recognition method based on a finite state machine is adopted to obtain conversation texts of customers and sales, and a pre-established finite state machine is used to determine the conversation state according to chronological order and specific words, distinguishing different customers in segments.
It realizes the ease of expansion of new customer identification, guaranteed accuracy, and has low dependence on labeled data, which facilitates the migration of business scenarios.
Smart Images
Figure CN113919364B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of text processing, and in particular to a new customer identification method and device based on a finite state machine. Background Art
[0002] In offline sales scenarios, collecting conversation data between salespeople and customers without infringing on customer privacy can help sellers iterate and upgrade their self-management and better serve customers. Data collection is generally done by recording conversation data through recording devices such as voice badge hardware worn by salespeople, and then processing, analyzing and summarizing conversation data in combination with AI technologies such as automatic speech recognition and natural language processing.
[0003] However, when the recording time is short and the voiceprint information is incomplete, it is difficult to distinguish different customers from a day or half a day of voice, especially when the cost of badge hardware needs to be controlled. How to distinguish customers from a whole voice text through natural language technology has become an urgent problem to be solved.
[0004] To address this problem, the existing technology trains machine learning models to perform binary classification on each sentence, that is, classifying it into a start sentence and a non-start sentence, so as to distinguish customers in the voice text. This method relies on the construction of feature engineering, which is time-consuming. At the same time, the accuracy is low in business scenarios. For example, in the cosmetics shopping guide scenario, there will be a large number of customers trying out the products. At this time, there will be waiting time, and there will also be waiting words like "hello" that strongly interfere with the binary classification model. Summary of the invention
[0005] In order to overcome the problems existing in the related art, the present invention provides a new customer identification method and device based on a finite state machine.
[0006] According to a first aspect of an embodiment of the present invention, a new customer identification method based on a finite state machine is provided, comprising:
[0007] Get the conversation text between customers and sales;
[0008] Inputting the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue;
[0009] The conversation text is segmented according to the determined states to distinguish different customers in the context of the conversation text.
[0010] According to a second aspect of an embodiment of the present invention, a new customer identification system based on a finite state machine is provided, comprising:
[0011] The dialogue acquisition module is used to obtain the dialogue text between customers and sales;
[0012] A state determination module, used to input the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue;
[0013] The conversation segmentation module is used to segment the conversation text according to the state determined by the state determination module to distinguish different customers in the context of the conversation text.
[0014] According to a third aspect of an embodiment of the present invention, a terminal device is provided, including:
[0015] Processor; and
[0016] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method as described above.
[0017] According to a fourth aspect of an embodiment of the present invention, there is provided a non-transitory machine-readable storage medium having executable code stored thereon, which, when executed by a processor of an electronic device, causes the processor to execute the method as described above.
[0018] The technical solution provided by the embodiments of the present invention may have the following beneficial effects:
[0019] It relies on a customized finite state machine to identify new customers. It is easy to expand, has guaranteed accuracy, has low dependence on labeled data, and facilitates the migration of business scenarios.
[0020] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.
[0022] Figure 1 is a flow chart of a new customer identification method based on a finite state machine according to an exemplary embodiment of the present invention;
[0023] Figure 2 is a schematic diagram of the principle of a finite state machine provided by an embodiment of the present invention;
[0024] Figure 3 is a structural block diagram of a new customer identification device based on a finite state machine according to an exemplary embodiment of the present invention;
[0025] Figure 4 It is a schematic diagram showing the structure of a computing device according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION
[0026] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.
[0027] The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0028] It should be understood that although the terms "first", "second", "third", etc. may be used to describe various information in the present invention, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.
[0029] The present invention realizes new customer identification based on a finite state machine. To facilitate the understanding of the readers, the relevant knowledge of the finite state machine is briefly described below.
[0030] Finite state machine is a tool used to model object behavior. Its main function is to describe the state sequence that an object goes through during its life cycle and how it responds to various events from the outside world. The state machine can be summarized into four elements, namely the current state, condition, action, and next state. "Current state" and "condition" are causes, and "action" and "next state" are results. The details are as follows:
[0031] ① Current state: refers to the current state.
[0032] ②Condition: Also known as "event". When a condition is met, an action will be triggered or a state transition will be performed.
[0033] ③Action: The action executed after the condition is met. After the action is executed, you can migrate to a new state or keep the original state. The action is not required. When the condition is met, you can also migrate to the new state directly without executing any action.
[0034] ④ Sub-state: The new state to be moved to after the conditions are met. "Sub-state" is relative to "current state". Once the "sub-state" is activated, it will transform into the new "current state".
[0035] The technical solutions of the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0036] Figure 1 The figure is a flow chart of a new customer identification method based on a finite state machine according to an exemplary embodiment of the present invention.
[0037] See also Figure 1 , the method comprising:
[0038] 110. Get the conversation text between customers and sales;
[0039] Specifically, the conversation between the customer and the salesperson can be recorded through the voice badge worn by the salesperson, and then converted into conversation text through ASR (Automatic Speech Recognition) to generate a conversation arranged in chronological order.
[0040] In addition, to facilitate subsequent analysis, it is necessary to first mark the sales conversations that belong to sales and the customer conversations that belong to customers in the conversation text. Specifically, the sales voice text can be collected in advance to train the voiceprint recognition model, and then the sales conversations can be identified and marked by the model, and other conversations can be uniformly marked as customer conversations.
[0041] 120. Inputting the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue;
[0042] Specifically, the purpose of this step is to determine whether each sales dialogue occurs when starting a dialogue with a new customer, so as to facilitate the subsequent classification of different customers in the dialogue text context.
[0043] Therefore, in the finite state machine, two states can be defined: starting with a new customer and starting with a non-new customer. Some characteristics of the sales conversation, such as the time when the conversation occurs and specific words contained in the conversation, can be used as conditions for state transition, thereby completing the definition of each element in the finite state machine.
[0044] 130. Segment the conversation text according to the determined status to distinguish different customers in the context of the conversation text.
[0045] Specifically, based on the state determined in step 120, the conversation text can be divided. Take a sales conversation extracted from a conversation text at a cosmetics counter as an example to illustrate as follows:
[0046] 13:46:57 Hello
[0047] 13:46:58 Welcome
[0048] 13:47:03 You want the oil, right?
[0049] 13:49:05 I'll give you a new one
[0050] 13:52:13Welcome to visit us next time
[0051] 13:54:29What kind of things are you looking for?
[0052] Through the finite state machine judgment, the first sentence "Hello" and the last sentence "What are you looking for?" belong to the "new customer start" state, that is, the first to fifth sentences belong to the conversation text between the salesperson and the previous customer, and the sixth sentence onwards is the conversation text between the salesperson and the next new customer. The judgment is made in sequence, thereby separating the conversation texts between the salesperson and different customers and realizing customer segmentation.
[0053] Optionally, in this embodiment, step 120 specifically includes:
[0054] 1201. Determine an event defined in a finite state machine corresponding to the sales dialogue according to the occurrence time of the sales dialogue and the target word therein;
[0055] Specifically, in this embodiment, the target words include: special words at the beginning of a conversation that express the semantics of cohesion, such as: then, just...; start words that express the semantics of the beginning of a conversation, such as: hello, hello, good morning, welcome...; end words that express the semantics of the end of a conversation, such as: goodbye, bye, welcome next time, take care...; middle words that are in the middle of a conversation and express the semantics of cohesion, such as: no, you are fine, for your own good... The above-mentioned vocabulary of various target words can be added as needed.
[0056] like Figure 2 As shown, the states defined in the finite state machine include a first state representing "new customer start" (i.e. Figure 2 State 1 in ) and the second state of "non-new customer start" (i.e. Figure 2 State 2).
[0057] The events in the finite state machine are configured to be defined according to the time interval between the occurrence of the two sales dialogues and the preset target words, including: the first event, the second event, the third event and the fourth event (respectively Figure 2 As shown in case 1, case 2, case 3 and case 4), the first event includes a first sub-event case 1-(1) and a second sub-event case 1-(2), the first sub-event is defined as the time interval between the current sales dialogue and the previous sales dialogue exceeds the first threshold, the second sub-event is defined as the time interval between the current sales dialogue and the previous sales dialogue exceeds the second threshold, but does not meet the second event; the second event is defined as the time interval between the current sales dialogue and the previous sales dialogue exceeds the second threshold, but does not exceed the first threshold, and the beginning of the current sales dialogue is a special word or there is a start word in the previous sales dialogue; the third event is defined as the time interval between the current sales dialogue and the previous sales dialogue does not exceed the second threshold, and meets one of the following two conditions: (1) the previous sales dialogue does not have a start word, and the current sales dialogue has a start word, (2) the current sales dialogue does not have a start word, and the previous sales dialogue has an end word; the fourth event is defined as the middle of the current sales dialogue contains a middle word.
[0058] Step 1201 specifically includes the following steps:
[0059] Determining whether the current sales dialogue satisfies the fourth event;
[0060] If the fourth event is not satisfied, determining whether the time interval between the current sales dialogue and the previous sales dialogue satisfies the first sub-event;
[0061] If the first sub-event is not satisfied, determining whether the current sales conversation satisfies the second event or the third event according to the time interval;
[0062] If the time interval satisfies the second event, further determining whether the text content of the sales dialogue satisfies the second event;
[0063] If the time interval satisfies the third event, further determining whether the text content of the sales dialogue satisfies the third event;
[0064] Repeat the above steps for each sales conversation in chronological order.
[0065] Specifically, the event judgment logic of the above finite state machine can be implemented using Python programming. The pseudo code logic is as follows:
[0066] Step 1: Determine whether case 4 is satisfied. If not, proceed to step 2
[0067] Step 2: Determine whether the time interval between the upper and lower sentences satisfies the first sub-event case 1-(1) in case 1. If not, select step 3 / step 4 according to the time interval.
[0068] Step 3: If the time interval between the upper and lower sentences meets case 2, determine whether the content meets case 2;
[0069] Step 4: If the time interval between the upper and lower sentences meets case 3, determine whether the content meets case 3;
[0070] Step 5: Loop through steps 1-4 for each sentence in chronological order.
[0071] Since the sales conversations are marked with the time of occurrence, the time interval between the sales conversations can be determined based on the time of occurrence of the sales conversations. Combined with the text content of the sales conversations, it can be determined whether the target words are included, so as to determine whether each sales conversation conforms to the corresponding event.
[0072] 1202. Determine an action defined in a finite state machine corresponding to the sales dialogue according to the determined event;
[0073] Specifically, step 1202 specifically includes:
[0074] If the current sales dialogue belongs to the first sub-event or the second sub-event in the first event, the sales dialogue remains in the first state;
[0075] If the current sales conversation belongs to the second event, the sales conversation is transferred from the first state to the second state;
[0076] If the current sales dialogue belongs to the third event, the sales dialogue is transferred from the second state to the first state;
[0077] If the current sales dialogue belongs to the fourth event, the sales dialogue remains in the first state.
[0078] 1203. Determine the state after the action is completed.
[0079] Specifically, after the action is executed, it can be determined whether the current sales dialogue corresponds to the first state of "new customer starting" or the second state of "non-new customer starting".
[0080] Optionally, in this embodiment, step 130 specifically includes:
[0081] The conversation texts belonging to different context states are segmented, and each conversation text is defined as a new customer in turn.
[0082] Specifically, after segmentation, each conversation text can be subjected to business analysis in terms of specific sales tactics, service quality, and other aspects.
[0083] Figure 3 is a structural block diagram of a new customer identification device based on a finite state machine according to an exemplary embodiment of the present invention.
[0084] See also Figure 3 , the device comprises:
[0085] The dialogue acquisition module is used to obtain the dialogue text between customers and sales;
[0086] A state determination module, used to input the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue;
[0087] The conversation segmentation module is used to segment the conversation text according to the state determined by the state determination module to distinguish different customers in the context of the conversation text.
[0088] Optionally, in this embodiment, the state determination module specifically includes:
[0089] An event determination unit, configured to determine an event defined in a finite state machine corresponding to the sales dialogue according to the occurrence time of the sales dialogue and the target word therein, wherein the event in the finite state machine is configured to be defined according to the time interval between the occurrence times of the upper and lower sales dialogues and the preset target word;
[0090] An action determination unit, used to determine the action defined in the finite state machine corresponding to the sales dialogue according to the determined event;
[0091] The state determination unit is used to determine the state after the action is executed, wherein the state defined in the finite state machine includes a first state representing "new customer starts" and a second state representing "non-new customer starts".
[0092] Optionally, in this embodiment, the target words include special words at the beginning of a conversation and indicating cohesive semantics, start words indicating conversation start semantics, end words indicating conversation end semantics, and middle words in the middle of a conversation and indicating cohesive semantics; the states defined in the finite state machine include a first state and a second state; the defined events include: a first event, a second event, a third event, and a fourth event, the first event includes a first sub-event and a second sub-event, the first sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds a first threshold, the second sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds a second threshold, but does not meet the second event; the second event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds the second threshold, but does not exceed the first threshold, and the beginning of the current sales conversation is a special word or the previous sales conversation has a start word; the third event is defined as the time interval between the current sales conversation and the previous sales conversation does not exceed the second threshold, and meets one of the following two conditions: (1) the previous sales conversation does not have a start word, and the current sales conversation has a start word, (2) the current sales conversation does not have a start word, and the previous sales conversation has an end word; the fourth event is defined as the middle of the current sales conversation contains a middle word;
[0093] The event determination unit is specifically used to:
[0094] Determining whether the current sales dialogue satisfies the fourth event;
[0095] If the fourth event is not satisfied, determining whether the time interval between the current sales dialogue and the previous sales dialogue satisfies the first sub-event;
[0096] If the first sub-event is not satisfied, determining whether the current sales conversation satisfies the second event or the third event according to the time interval;
[0097] If the time interval satisfies the second event, further determining whether the text content of the sales dialogue satisfies the second event;
[0098] If the time interval satisfies the third event, further determining whether the text content of the sales dialogue satisfies the third event;
[0099] Repeat the above steps for each sales conversation in chronological order.
[0100] Optionally, in this embodiment, the action determination unit is specifically configured to:
[0101] If the current sales dialogue belongs to the first sub-event or the second sub-event in the first event, the sales dialogue remains in the first state;
[0102] If the current sales conversation belongs to the second event, the sales conversation is transferred from the first state to the second state;
[0103] If the current sales dialogue belongs to the third event, the sales dialogue is transferred from the second state to the first state;
[0104] If the current sales dialogue belongs to the fourth event, the sales dialogue remains in the first state.
[0105] Optionally, in this embodiment, the conversation segmentation module is specifically used to:
[0106] The conversation texts belonging to different context states are segmented, and each conversation text is defined as a new customer in turn.
[0107] Regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.
[0108] Figure 4 It is a schematic diagram showing the structure of a computing device according to an exemplary embodiment of the present invention.
[0109] See also Figure 4 , the computing device 400 includes a memory 410 and a processor 420 .
[0110] The processor 420 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
[0111] The memory 410 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. Among them, ROM can store static data or instructions required by the processor 420 or other modules of the computer. The permanent storage device may be a readable and writable storage device. The permanent storage device may be a non-volatile storage device that does not lose the stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory may store some or all instructions and data required by the processor at run time. In addition, the memory 410 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 410 may include a readable and / or writable removable storage device, such as a laser disc (CD), a read-only digital versatile disc (such as a DVD-ROM, a double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a mini SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer-readable storage medium does not include carrier waves and transient electronic signals transmitted wirelessly or wired.
[0112] The memory 410 stores executable codes, and when the executable codes are processed by the processor 420 , the processor 420 can execute part or all of the methods described above.
[0113] In addition, the method according to the present invention may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing part or all of the steps in the above method of the present invention.
[0114] Alternatively, the present invention may also be implemented as a non-temporary machine-readable storage medium (or computer-readable storage medium, or machine-readable storage medium) on which executable code (or computer program, or computer instruction code) is stored. When the executable code (or computer program, or computer instruction code) is executed by a processor of an electronic device (or computing device, server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present invention.
[0115] The scheme of the present invention has been described in detail above with reference to the accompanying drawings. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. Those skilled in the art should also be aware that the actions and modules involved in the specification are not necessarily required by the present invention. In addition, it can be understood that the steps in the method of the embodiment of the present invention can be adjusted in order, merged and deleted according to actual needs, and the modules in the device of the embodiment of the present invention can be merged, divided and deleted according to actual needs.
[0116] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the disclosure herein may be implemented as electronic hardware, computer software, or a combination of both.
[0117] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system and method according to multiple embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or code, and the part of the module, program segment or code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0118] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terms used herein are selected to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.
Claims
1. A new customer identification method based on finite state machine, characterized in that: include: Get the conversation text between customers and sales; Inputting the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue, specifically comprising: determining an event defined in the finite state machine corresponding to the sales dialogue according to the occurrence time and text content of the sales dialogue, wherein the event in the finite state machine is configured to be defined according to the time interval between the occurrence times of the upper and lower sales dialogues and a preset target word; determining an action defined in the finite state machine corresponding to the sales dialogue according to the determined event; determining the state after the action is executed, wherein the state defined in the finite state machine includes a first state indicating "new customer starts" and a second state indicating "non-new customer starts"; Segmenting the conversation text according to the determined states to distinguish different customers in the context of the conversation text; The target words include special words at the beginning of a conversation and indicating the semantics of cohesion, start words indicating the semantics of the beginning of a conversation, end words indicating the semantics of the end of a conversation, and middle words in the middle of a conversation and indicating the semantics of cohesion; the states defined in the finite state machine include a first state and a second state; the defined events include a first event, a second event, a third event, and a fourth event, the first event includes a first sub-event and a second sub-event, the first sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds a first threshold, the second sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds a second threshold, but does not meet the second event; the second event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds the second threshold, but does not exceed the first threshold, and the beginning of the current sales conversation is a special word or the previous sales conversation contains a start word; the third event is defined as the time interval between the current sales conversation and the previous sales conversation does not exceed the second threshold, and meets one of the following two conditions: (1) The previous sales dialogue has no start word, and the current sales dialogue has a start word; (2) the current sales dialogue has no start word, and the previous sales dialogue has an end word; the fourth event is defined as the middle of the current sales dialogue containing a middle word; the event defined in the finite state machine corresponding to the sales dialogue is determined according to the occurrence time of the sales dialogue and the target word therein, specifically including: judging whether the current sales dialogue satisfies the fourth event; if the fourth event is not satisfied, judging whether the time interval between the current sales dialogue and the previous sales dialogue satisfies the first sub-event; if the first sub-event is not satisfied, judging whether the current sales dialogue satisfies the second event or the third event according to the time interval; if the time interval satisfies the second event, further judging whether the text content of the sales dialogue satisfies the second event; if the time interval satisfies the third event, further judging whether the text content of the sales dialogue satisfies the third event; and looping through each sales dialogue in chronological order.
2. The method according to claim 1, characterized in that Determining the action defined in the finite state machine corresponding to the sales dialogue according to the determined event specifically includes: If the current sales dialogue belongs to the first sub-event or the second sub-event in the first event, the sales dialogue remains in the first state; If the current sales conversation belongs to the second event, the sales conversation is transferred from the first state to the second state; If the current sales dialogue belongs to the third event, the sales dialogue is transferred from the second state to the first state; If the current sales dialogue belongs to the fourth event, the sales dialogue remains in the first state.
3. A new customer identification device based on a finite state machine, characterized in that: include: The dialogue acquisition module is used to obtain the dialogue text between customers and sales; A state determination module, used to input the sales dialogue in the dialogue text into a pre-established finite state machine in chronological order to determine the state corresponding to the sales dialogue; A conversation segmentation module, used for segmenting the conversation text according to the state determined by the state determination module, so as to distinguish different customers in the context of the conversation text; The state determination module specifically includes: an event determination unit, which is used to determine an event defined in a finite state machine corresponding to the sales dialogue according to the occurrence time of the sales dialogue and the target word therein, wherein the event in the finite state machine is configured to be defined according to the time interval between the occurrence times of the upper and lower sales dialogues and the preset target word; an action determination unit, which is used to determine an action defined in the finite state machine corresponding to the sales dialogue according to the determined event; a state determination unit, which is used to determine the state after the action is executed, wherein the state defined in the finite state machine includes a first state indicating "new customer starts" and a second state indicating "non-new customer starts"; The target words include special words at the beginning of the conversation and indicating the semantics of cohesion, start words indicating the semantics of the beginning of the conversation, end words indicating the semantics of the end of the conversation, and middle words in the middle of the conversation and indicating the semantics of cohesion; the states defined in the finite state machine include the first state and the second state; the defined events include: the first event, the second event, the third event, and the fourth event, the first event includes the first sub-event and the second sub-event, the first sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds the first threshold, the second sub-event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds the second threshold, but does not meet the second event; the second event is defined as the time interval between the current sales conversation and the previous sales conversation exceeds the second threshold, but does not exceed the first threshold, and the beginning of the current sales conversation is a special word or there is a start word in the previous sales conversation; the third event is defined as the time interval between the current sales conversation and the previous sales conversation does not exceed the second threshold, And it meets one of the following two conditions: (1) the previous sales dialogue has no start word, and the current sales dialogue has a start word, (2) the current sales dialogue has no start word, and the previous sales dialogue has an end word; the fourth event is defined as the middle of the current sales dialogue containing a middle word; the event determination unit is specifically used to: determine whether the current sales dialogue satisfies the fourth event; if the fourth event is not satisfied, determine whether the time interval between the current sales dialogue and the previous sales dialogue satisfies the first sub-event; if the first sub-event is not satisfied, determine whether the current sales dialogue satisfies the second event or the third event based on the time interval; if the time interval satisfies the second event, further determine whether the text content of the sales dialogue satisfies the second event; if the time interval satisfies the third event, further determine whether the text content of the sales dialogue satisfies the third event; and loop through each sales dialogue in chronological order.
4. The device according to claim 3, characterized in that The action determination unit is specifically used for: If the current sales dialogue belongs to the first sub-event or the second sub-event in the first event, the sales dialogue remains in the first state; If the current sales conversation belongs to the second event, the sales conversation is transferred from the first state to the second state; If the current sales dialogue belongs to the third event, the sales dialogue is transferred from the second state to the first state; If the current sales dialogue belongs to the fourth event, the sales dialogue remains in the first state.
5. A terminal device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 2.
6. A non-transitory machine-readable storage medium having executable code stored thereon, characterized in that: When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method as claimed in any one of claims 1 to 2.