Method and device for extracting keywords from feedback information

Abstract

The present invention discloses a method and device for extracting keywords of feedback information, and relates to the field of computer technology. A specific implementation of the method includes: performing word segmentation processing on feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information; using a word frequency-inverse document frequency algorithm to calculate the weight value of at least one target word segmentation in the feedback information; and confirming the keywords of the feedback information according to the weight value of at least one target word segmentation and a preset weight value. This implementation utilizes a preset word segmentation processing algorithm to obtain the target word segmentation of the feedback information, and then confirms the keywords of the feedback information by the weight value of the target word segmentation, thereby improving the accuracy of keyword extraction, greatly obtaining valuable feedback information, reducing the probability of important content being ignored, and improving work efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for extracting keywords from feedback information. Background Art

[0002] In today's society, competition is increasingly fierce. All walks of life make decisions based on user feedback and complete product updates and improvements, so that their products can have advantages and improve user experience. Before processing feedback information, it is necessary to classify the feedback information, for example, determine whether the feedback information belongs to the business or technical category.

[0003] At present, feedback information is screened and classified manually. This method is laborious when there is a lot of feedback information, consumes a lot of manpower, and is easy to miss important information. Considering that the keywords in the feedback information can be used to classify the feedback information, how to extract the keywords in the feedback information is of great significance. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a method and device for extracting keywords from feedback information, which can improve the accuracy of keyword extraction, greatly obtain valuable feedback information, reduce the probability of important content being ignored, and improve work efficiency.

[0005] To achieve the above objective, according to one aspect of an embodiment of the present invention, a method for extracting keywords from feedback information is provided.

[0006] A method for extracting keywords from feedback information according to an embodiment of the present invention includes: performing word segmentation processing on feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information; calculating a weight value of the at least one target word segmentation in the feedback information using a word frequency-inverse document frequency algorithm; and confirming keywords of the feedback information based on the weight value of the at least one target word segmentation and a preset weight value.

[0007] Optionally, performing word segmentation processing on the feedback information based on a custom word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information includes: performing word segmentation processing on the feedback information based on an improved string matching algorithm to obtain original word segmentations of the feedback information; performing optimized word segmentation processing on the original word segmentations based on a multiple hidden Markov model optimized word segmentation algorithm to obtain the at least one target word segmentation.

[0008] Optionally, the improved string matching algorithm includes: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule.

[0009] Optionally, the multiple hidden Markov model optimized word segmentation algorithm includes: adding multiple semantics and noise information to the hidden Markov model to form an optimized word segmentation algorithm.

[0010] Optionally, confirming the keywords of the feedback information according to the weight value of the at least one target participle and the preset weight value includes: for each target participle in the at least one target participle, judging whether the weight value of the target participle is greater than the preset weight value, and if so, taking the target participle as the keyword; if the weight values ​​of all target participles in the at least one target participle are not greater than the preset weight value, obtaining the maximum value of the weight value of the at least one target participle, and taking the target participle corresponding to the maximum value as the keyword.

[0011] To achieve the above objective, according to another aspect of an embodiment of the present invention, a device for extracting keywords from feedback information is provided.

[0012] An apparatus for extracting keywords from feedback information according to an embodiment of the present invention comprises: an acquisition module, configured to perform word segmentation processing on feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information; a calculation module, configured to calculate a weight value of the at least one target word segmentation in the feedback information by using a word frequency-inverse document frequency algorithm; and a confirmation module, configured to confirm keywords of the feedback information according to the weight value of the at least one target word segmentation and a preset weight value.

[0013] Optionally, the acquisition module is also used to: perform word segmentation processing on the feedback information based on an improved string matching algorithm to obtain original word segments of the feedback information; perform optimized word segmentation processing on the original word segments based on a multiple hidden Markov model optimized word segmentation algorithm to obtain at least one target word segmentation.

[0014] Optionally, the improved string matching algorithm includes: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule.

[0015] Optionally, the multiple hidden Markov model optimized word segmentation algorithm includes: adding multiple semantics and noise information to the hidden Markov model to form an optimized word segmentation algorithm.

[0016] Optionally, the confirmation module is also used to: for each target participle in the at least one target participle, determine whether the weight value of the target participle is greater than the preset weight value, and if so, use the target participle as a keyword; if the weight values ​​of all target participles in the at least one target participle are not greater than the preset weight value, obtain the maximum value of the weight value of the at least one target participle, and use the target participle corresponding to the maximum value as a keyword.

[0017] To achieve the above objective, according to another aspect of an embodiment of the present invention, an electronic device is provided.

[0018] An electronic device according to an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the keyword extraction method according to the embodiment of the present invention.

[0019] To achieve the above objective, according to another aspect of an embodiment of the present invention, a computer readable medium is provided.

[0020] A computer-readable medium according to an embodiment of the present invention stores a computer program, and when the program is executed by a processor, the method for extracting keywords according to an embodiment of the present invention is implemented.

[0021] An embodiment of the above invention has the following advantages or beneficial effects: it is possible to use a preset word segmentation processing algorithm to obtain the target word segmentation of the feedback information, and then confirm the keyword of the feedback information by the weight value of the target word segmentation, thereby improving the accuracy of keyword extraction, greatly obtaining valuable feedback information, reducing the probability of important content being ignored, and improving work efficiency; when the feedback information is segmented in the embodiment of the present invention, the original word segmentation is obtained by using an improved string matching algorithm, which has the characteristics of easy implementation and relatively fast word segmentation speed; in the embodiment of the present invention, a multiple hidden Markov model optimization word segmentation algorithm is used to optimize the original word segmentation, and multiple semantics and noise information are added to the model, so that the feedback information can be segmented in combination with the semantic context, solving the cross ambiguity and combination ambiguity steady state, and improving the accuracy of word segmentation; when confirming the keyword in the embodiment of the present invention, it is judged whether the target word segmentation is a keyword according to the preset weight value, and when the weight value of the target word segmentation is not greater than the preset weight value, the target word segmentation corresponding to the maximum weight value is selected as the keyword, thereby improving the practicality of the scheme.

[0022] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0024] Figure 1 is a schematic diagram of main steps of a method for extracting keywords from feedback information according to an embodiment of the present invention;

[0025] Figure 2 is a schematic diagram of the main process of a method for extracting keywords from feedback information according to a reference embodiment of the present invention;

[0026] Figure 3 is a schematic diagram of main modules of an apparatus for extracting keywords from feedback information according to an embodiment of the present invention;

[0027] Figure 4 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0028] Figure 5 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0030] Figure 1 Schematic diagram of the main steps of the method for extracting keywords from feedback information according to an embodiment of the present invention. Figure 1 As shown, the main steps of the method for extracting keywords from feedback information in the embodiment of the present invention may include:

[0031] Step S101: Segment the feedback information based on a preset segmentation algorithm to obtain at least one target segmentation of the feedback information. The feedback information in the present invention refers to the reaction information expressed by the user through voice or text after purchasing a product, consuming a product, or enjoying a product service. In the present invention, it is necessary to segment the feedback information and extract the keywords of the feedback information. Considering that this scheme is a keyword extraction for the particularity of Chinese, the feedback information in this scheme can be Chinese text information, or it can be Chinese text information converted from voice information through a voice-to-text tool. Of course, it can also be Chinese text information converted from other forms of information, and there is no limitation on this.

[0032] As another reference embodiment of the present invention, the above step S101 performs word segmentation processing on the feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information, which may include: performing word segmentation processing on the feedback information based on an improved string matching algorithm to obtain the original word segmentation of the feedback information; performing optimized word segmentation processing on the original word segmentation based on a multiple hidden Markov model optimization word segmentation algorithm to obtain at least one target word segmentation. The string matching-based algorithm is also called a mechanical word segmentation method or a dictionary matching method. It mainly relies on the information of the dictionary, but does not use rule knowledge and statistical information. According to a certain strategy, the Chinese character string to be segmented is matched one by one with the entries in the dictionary. If the entry is found in the dictionary, the match is successful, otherwise other corresponding processing is performed. The Hidden Markov Model (HMM) is a statistical model that is used to describe a Markov process containing implicit unknown parameters.

[0033] In an embodiment of the present invention, an improved string matching algorithm may include: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule. The minimum granularity segmentation rule refers to a rule for segmenting an original string into an indivisible sequence, and the maximum matching rule is divided into a forward maximum matching algorithm and a reverse maximum matching algorithm according to different scanning directions. Among them, the basic principle of the forward maximum matching algorithm is: select a symbol string of a fixed length of Chinese characters as the maximum symbol string, match the maximum symbol string with a word entry in the dictionary, and if it cannot be matched, remove a Chinese character and continue matching until the corresponding word is found in the dictionary. Taking the forward maximum matching algorithm where the matching direction is from left to right and the subtraction direction is from right to left as an example, the specific matching process is as follows:

[0034] (1) Initialize the string and set the maximum symbol string length. Let H1 be the string to be analyzed, whose initial value is feedback information, and H2 be the segmentation result string, whose initial value is empty;

[0035] (2) If H1 is not empty, take the candidate substring M from the left side of H1; if H1 is empty, output H2 as the word segmentation result;

[0036] (3) Compare M with the dictionary. If M is in the dictionary's vocabulary, add M to H2 and remove M from H1, then jump to step (2). If M is not in the vocabulary, remove the rightmost word of M, then jump to step (3).

[0037] The reverse maximum matching algorithm is similar to the forward maximum matching algorithm, except that the matching direction is from right to left and the word reduction direction is from left to right. The forward maximum matching algorithm and the reverse maximum matching algorithm have the advantages of being easy to implement and having a relatively fast word segmentation speed. However, due to the profoundness and extensiveness of the Chinese language, string matching algorithms do not have good solutions for cross ambiguity and combinatorial ambiguity. Cross ambiguity, for example, when segmenting the phrase "in the mind", since both "mind" and "sea" are words, this phrase can be segmented into "mind" and "in" or "brain" and "in the sea"; combinatorial ambiguity must be judged based on the entire sentence. For example, in the sentence "He has extraordinary talents", "talents" is a word, but in the sentence "Only he is capable of taking up this position", "talents" is not a word. Both cross ambiguity and combinatorial ambiguity are due to the lack of human knowledge for understanding, and it is difficult for a computer to know which solution is correct. It is necessary to rely on the language context of the context to give a correct segmentation. Therefore, after using the string matching algorithm to perform word segmentation on the feedback information, the present invention performs secondary word segmentation through a multiple hidden Markov model.

[0038] In an embodiment of the present invention, the multiple hidden Markov model optimization word segmentation algorithm may include: adding multiple semantics and noise information to the hidden Markov model to form an optimized word segmentation algorithm. Multiple semantics refers to different understanding scenarios of the same text information. For example, "we are a group of three" can be understood as "we / a group of three" or "we are three / a group". For example, "the demeanor of a poet" may refer to a poet or not. Noise information refers to external conditions that affect the final word segmentation result. The multiple hidden Markov model optimization word segmentation algorithm of the present invention is to add external conditions that affect the final result, such as multiple semantics and noise information of Chinese, to the model to form a system model. In fact, it is a combination of multiple HMMs, and each layer of HMM adopts the N-Best strategy to send the best N results generated to the word graph for use by a higher-level model. The N-Best strategy refers to the N shortest path strategy. By adopting the N-best method, the N best rough segmentation results can be quickly obtained. Then, on this result set, the hidden Markov algorithm is used to identify specific information in the feedback information, such as business details, search, and nested professional vocabulary. In addition, the multiple hidden Markov models in this solution construct different data sources for multiple semantics without performing other processing such as excision, that is, each semantic is treated as a raw data, reducing the coupling between data. Step S102: Calculate the weight value of at least one target word in the feedback information using the term frequency-inverse document frequency algorithm. The term frequency-inverse document frequency algorithm (TF-IDF algorithm for short) is a commonly used weighting technology for information retrieval and data mining. The TF-IDF value (i.e., weight value) of a word can be calculated according to the TF-IDF algorithm. The more important a word is to the article, the greater its TF-IDF value.

[0039] The TF-IDF value of the target word is: TF*IDF. First, calculate the term frequency (TF) of the target word in the feedback information, that is, the frequency of the target word in the feedback information. In order to prevent it from being biased towards long files, the TF value is normalized to the number of words. The specific calculation formula is as follows:

[0040]

[0041] From the above formula, we can see that there are k target words in the feedback information P, n i,P refers to the target participle x i The number of occurrences in the feedback information P, ∑ k n k,PIt refers to the sum of the number of times k target words appear in the feedback information P. Inverse Document Frequency (IDF) is a measure of the general importance of a word. The IDF of a specific word can be obtained by dividing the total number of documents by the number of documents containing the word, and then taking the logarithm of the quotient. The TF-IDF value of the target word refers to the product of the TF value of the target word and the IDF value of the target word. The high word frequency in a specific file and the low file frequency of the word in the entire file collection can produce a high-weight TF-IDF. Therefore, TF-IDF can filter out common words and retain important words. The specific calculation formula of TF-IDF is as follows:

[0042]

[0043] In the above formula, TF-IDF(x i ,p) refers to the target participle x i The TF-IDF value in the feedback information P, TF(x i ,p) refers to the target participle x i The word frequency in the feedback information P, a is the total number of all feedback information, DF(x i ) is the target word x i The word frequency in all feedback information, k is the number of target segmented words contained in the feedback information P. Through the above formula, the TF-IDF value of the target segmented word in the feedback information can be calculated, that is, the weight value of the target segmented word in the feedback information.

[0044] Step S103: Confirm the keyword of the feedback information according to the weight value of at least one target segmentation and the preset weight value. The present invention determines whether the target segmentation is a keyword of the feedback information according to the weight value of the target segmentation. If so, the feedback information is classified by the keyword, which can greatly obtain valuable feedback information, reduce the probability of important content being ignored, and improve work efficiency.

[0045] As another reference embodiment of the present invention, step S103 confirms the keywords of the feedback information according to the weight value and the preset weight value of at least one target participle, which may include: for each target participle in at least one target participle, judging whether the weight value of the target participle is greater than the preset weight value, if so, the target participle is considered to be a keyword; when the weight values ​​of all target participles in at least one target participle are not greater than the preset weight value, obtaining the maximum value of the weight value of at least one target participle, and considering the target participle corresponding to the maximum value to be a keyword. This step specifically explains how to use the weight value of the target participle to judge whether the target participle is a keyword, and sets that when the weight values ​​of all target participles in the feedback information are less than the preset weight value, the target participle with the largest weight value is directly selected as the keyword.

[0046] Figure 2 FIG. 1 is a schematic diagram of the main process of a method for extracting keywords from feedback information according to a reference embodiment of the present invention. Figure 2 As shown, the main process of the method for extracting keywords from feedback information in an embodiment of the present invention may include:

[0047] Step S201: performing word segmentation processing on the feedback information P based on the improved string matching algorithm to obtain the original word segmentation of the feedback information P;

[0048] Step S202: performing optimized word segmentation processing on the original word segmentation based on the multiple hidden Markov model optimized word segmentation algorithm to obtain at least one target word segmentation of the feedback information P;

[0049] Step S203: Select any one of the acquired target segmentations as the target segmentation A, and calculate the weight value of the target segmentation A in the feedback information P using the TF-IDF algorithm;

[0050] Step S204: determine whether the weight value of the target word A in the feedback information P is greater than the preset weight value, if so, execute step S205, if not, execute step S206;

[0051] Step S205: confirm that the target segmented word A is a keyword of the feedback information P, and put it into the keyword set S of the feedback information P;

[0052] Step S206: Determine whether to analyze all target words of the feedback information P, if yes, execute step S207;

[0053] Step S207: Determine whether the keyword set S is empty, if so, execute step S208, if not, execute step S210;

[0054] Step S208: Obtain the maximum value of the weight value of the target word of the feedback information P;

[0055] Step S209: confirm that the target segmentation corresponding to the maximum weight value is the keyword of the feedback information P, and put the keyword into the keyword set S;

[0056] Step S210: Return the keyword set S of the feedback information P.

[0057] The improved string matching algorithm in step S201 and the multiple hidden Markov model optimized word segmentation algorithm in step S202 have been specifically introduced above and will not be repeated here. After returning the keyword set S of the feedback information P, the present solution can use the keywords in the keyword set S to analyze the classification of the feedback information P.

[0058] It can be seen from the technical solution for extracting keywords from feedback information according to the embodiment of the present invention that the preset word segmentation processing algorithm can be used to obtain the target word segmentation of the feedback information, and then the keyword of the feedback information can be confirmed by the weight value of the target word segmentation, thereby improving the accuracy of keyword extraction, greatly obtaining valuable feedback information, reducing the probability of important content being ignored, and improving work efficiency; when the feedback information is segmented in the embodiment of the present invention, the original word segmentation is obtained by using the improved string matching algorithm, which has the characteristics of easy implementation and relatively fast word segmentation speed; in the embodiment of the present invention, the original word segmentation is optimized by using the multiple hidden Markov model optimization word segmentation algorithm, and multiple semantics and noise information are added to the model, so that the feedback information can be segmented in combination with the semantic context, which solves the cross ambiguity and combination ambiguity steady state and improves the accuracy of word segmentation; when confirming the keyword in the embodiment of the present invention, it is judged whether the target word segmentation is a keyword according to the preset weight value, and when the weight value of the target word segmentation is not greater than the preset weight value, the target word segmentation corresponding to the maximum weight value is selected as the keyword, thereby improving the practicality of the scheme.

[0059] Figure 3 Schematic diagram of main modules of an apparatus for extracting keywords from feedback information according to an embodiment of the present invention. Figure 3 As shown, the apparatus 300 for extracting keywords from feedback information according to the embodiment of the present invention mainly includes the following modules: an acquisition module 301 , a calculation module 302 and a confirmation module 303 .

[0060] In the embodiment of the present invention, the acquisition module 301 can be used to perform word segmentation processing on the feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information. The calculation module 302 can be used to calculate the weight value of at least one target word segmentation in the feedback information using a word frequency-inverse document frequency algorithm. The confirmation module 303 can be used to confirm the keywords of the feedback information based on the weight value of at least one target word segmentation and the preset weight value.

[0061] In the embodiment of the present invention, the acquisition module 301 can also be used to: perform word segmentation processing on the feedback information based on the improved string matching algorithm to obtain the original word segmentation of the feedback information; perform optimized word segmentation processing on the original word segmentation based on the multiple hidden Markov model optimized word segmentation algorithm to obtain at least one target word segmentation.

[0062] In the embodiment of the present invention, the improved string matching algorithm may include: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule.

[0063] In the embodiment of the present invention, the multiple hidden Markov model optimization word segmentation algorithm may include: adding multiple semantics and noise information into the hidden Markov model to form an optimized word segmentation algorithm.

[0064] In an embodiment of the present invention, the confirmation module 303 can also be used to: for each target participle in at least one target participle, determine whether the weight value of the target participle is greater than a preset weight value, and if so, use the target participle as a keyword; if the weight values ​​of all target participles in at least one target participle are not greater than the preset weight value, obtain the maximum value of the weight value of at least one target participle, and use the target participle corresponding to the maximum value as a keyword.

[0065] From the above description, it can be seen that the preset word segmentation processing algorithm can be used to obtain the target word segmentation of the feedback information, and then the keyword of the feedback information can be confirmed by the weight value of the target word segmentation, so that the accuracy of keyword extraction can be improved, valuable feedback information can be greatly obtained, the probability of important content being ignored can be reduced, and work efficiency can be improved; when the feedback information is segmented in the embodiment of the present invention, the original word segmentation is obtained by using the improved string matching algorithm, which has the characteristics of easy implementation and relatively fast word segmentation speed; in the embodiment of the present invention, a multiple hidden Markov model optimization word segmentation algorithm is used to optimize the original word segmentation, and multiple semantics and noise information are added to the model, so that the feedback information can be segmented in combination with the semantic context, which solves the cross ambiguity and combination ambiguity steady state and improves the accuracy of word segmentation; when confirming the keyword in the embodiment of the present invention, it is judged whether the target word segmentation is a keyword according to the preset weight value, and when the weight value of the target word segmentation is not greater than the preset weight value, the target word segmentation corresponding to the maximum weight value is selected as the keyword, so that the practicality of this solution can be improved.

[0066] Figure 4 An exemplary system architecture 400 is shown to which the method for extracting keywords from feedback information or the apparatus for extracting keywords from feedback information according to the embodiments of the present invention can be applied.

[0067] like Figure 4As shown, system architecture 400 may include terminal devices 401, 402, 403, network 404 and server 405. Network 404 is used to provide a medium for communication links between terminal devices 401, 402, 403 and server 405. Network 404 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0068] Users can use terminal devices 401, 402, 403 to interact with server 405 through network 404 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 401, 402, 403, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0069] The terminal devices 401 , 402 , and 403 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0070] The server 405 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 401, 402, and 403. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as target push information, product information - only an example) to the terminal device.

[0071] It should be noted that the method for extracting keywords from feedback information provided in the embodiment of the present invention is generally executed by the server 405 , and accordingly, the device for extracting keywords from feedback information is generally disposed in the server 405 .

[0072] It should be understood that Figure 4 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0073] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 of a terminal device suitable for implementing an embodiment of the present invention. Figure 5 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0074] like Figure 5As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the system 500 are also stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0075] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed.

[0076] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above-mentioned functions defined in the system of the present invention are executed.

[0077] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0078] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a 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 box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession 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 box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0079] The modules involved in the embodiments of the present invention may be implemented in software or hardware. The modules described may also be set in a processor, for example, they may be described as: a processor includes an acquisition module, a calculation module and a confirmation module. The names of these units do not constitute a limitation on the module itself in some cases. For example, the acquisition module may also be described as "a module that performs word segmentation processing on feedback information based on a preset word segmentation processing algorithm and obtains at least one target word segmentation of the feedback information".

[0080] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: performing word segmentation processing on the feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information; using the word frequency-inverse document frequency algorithm, calculating the weight value of at least one target word segmentation in the feedback information; confirming the keywords of the feedback information according to the weight value of at least one target word segmentation and the preset weight value.

[0081] According to the technical solution of the embodiment of the present invention, the preset word segmentation processing algorithm can be used to obtain the target word segmentation of the feedback information, and then the keyword of the feedback information can be confirmed by the weight value of the target word segmentation, so as to improve the accuracy of keyword extraction, greatly obtain valuable feedback information, reduce the probability of important content being ignored, and improve work efficiency; when the feedback information is segmented in the embodiment of the present invention, the original word segmentation is obtained by using the improved string matching algorithm, which has the characteristics of easy implementation and relatively fast word segmentation speed; in the embodiment of the present invention, the multiple hidden Markov model optimization word segmentation algorithm is used to optimize the original word segmentation, and multiple semantics and noise information are added to the model, so that the feedback information can be segmented in combination with the semantic context, which solves the cross ambiguity and combination ambiguity steady state and improves the accuracy of word segmentation; when confirming the keyword in the embodiment of the present invention, it is judged whether the target word segmentation is a keyword according to the preset weight value, and when the weight value of the target word segmentation is not greater than the preset weight value, the target word segmentation corresponding to the maximum weight value is selected as the keyword, so as to improve the practicality of the scheme.

[0082] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for extracting keywords from feedback information, characterized in that: include: The feedback information is segmented based on a preset segmentation processing algorithm to obtain at least one target segmentation of the feedback information, including: performing segmentation on the feedback information based on an improved string matching algorithm to obtain the original segmentation of the feedback information; performing optimized segmentation on the original segmentation based on a multiple hidden Markov model optimized segmentation algorithm to obtain the at least one target segmentation; wherein the multiple hidden Markov model optimized segmentation algorithm includes: adding multiple semantics and noise information to the hidden Markov model to form an optimized segmentation algorithm; Calculate the weight value of the at least one target word segmentation in the feedback information by using a word frequency-inverse document frequency algorithm; The keyword of the feedback information is confirmed according to the weight value of the at least one target word segment and the preset weight value.

2. The method according to claim 1, characterized in that The improved string matching algorithm includes: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule.

3. The method according to claim 1, characterized in that: Confirming the keywords of the feedback information according to the weight value of the at least one target word segment and the preset weight value includes: For each target participle of the at least one target participle, determining whether the weight value of the target participle is greater than the preset weight value, and if so, using the target participle as a keyword; If the weight values ​​of all target participles in the at least one target participle are not greater than the preset weight value, the maximum value of the weight value of the at least one target participle is obtained, and the target participle corresponding to the maximum value is used as the keyword.

4. A device for extracting keywords from feedback information, characterized in that: include: The acquisition module is used to perform word segmentation processing on the feedback information based on a preset word segmentation processing algorithm to obtain at least one target word segmentation of the feedback information, including: performing word segmentation processing on the feedback information based on an improved string matching algorithm to obtain the original word segmentation of the feedback information; performing optimized word segmentation processing on the original word segmentation based on a multiple hidden Markov model optimized word segmentation algorithm to obtain the at least one target word segmentation; wherein the multiple hidden Markov model optimized word segmentation algorithm includes: adding multiple semantics and noise information to the hidden Markov model to form an optimized word segmentation algorithm; A calculation module, used to calculate a weight value of the at least one target word segmentation in the feedback information by using a word frequency-inverse document frequency algorithm; A confirmation module is used to confirm the keywords of the feedback information according to the weight value of the at least one target word segment and a preset weight value.

5. The device according to claim 4, characterized in that The improved string matching algorithm includes: a string matching algorithm using a minimum granularity segmentation rule and a maximum matching rule.

6. The device according to claim 4, characterized in that The confirmation module is also used for: For each target participle of the at least one target participle, determining whether the weight value of the target participle is greater than the preset weight value, and if so, using the target participle as a keyword; If the weight values ​​of all target participles in the at least one target participle are not greater than the preset weight value, the maximum value of the weight value of the at least one target participle is obtained, and the target participle corresponding to the maximum value is used as the keyword.

7. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 3.

8. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

Images