Comment analysis system and comment analysis method

Abstract

The invention provides a comment analysis system and a comment analysis method. The comment analysis system comprises a crawler module, a character detection module, an analysis module and a judgment module. And the crawler module is used for searching a plurality of user comments from the network database and storing the plurality of user comments in the database. The character detection module comprises a character library. The word stock is provided with a plurality of preset word strings. The text detection module is configured to collect at least one of the plurality of user comments including at least one of the plurality of word strings. The analysis module is used for analyzing at least one of the user comments. The judgment module is used for judging whether the word string has constructive significance in the at least one of the plurality of user comments or not according to the analysis result of the at least one of the plurality of user comments.

Description

Technical Field

[0001] The present disclosure relates to an automated data analysis technology, and in particular to a review analysis system and a review analysis method. Background Art

[0002] Generally speaking, the comments provided by users after purchasing a product can let product manufacturers or product sellers know the user's evaluation of the product and the potential defects of the product. However, with the increase in purchasing users, the product review information provided by users has also grown explosively, so it is quite impractical to manually collect relevant information. Summary of the invention

[0003] The present invention is directed to a comment analysis system and a comment analysis method, which can automatically crawl and analyze comment information in a network database.

[0004] According to an embodiment of the present disclosure, the comment analysis system of the present disclosure includes a crawler module, a text detection module, an analysis module and a determination module. The crawler module is coupled to a network database, and is used to search for multiple user comments from the network database, and store the multiple user comments in the database. The text detection module is coupled to the crawler module, and includes a character library. The character library has multiple pre-set character strings. The text detection module is used to collect at least one of the multiple user comments including at least one of the multiple character strings. The analysis module is coupled to the text detection module, and is used to analyze at least one of the multiple user comments. The determination module is coupled to the analysis module, and is used to determine whether the character string has constructive meaning in at least one of the multiple user comments based on the analysis result of at least one of the multiple user comments.

[0005] According to an embodiment of the present disclosure, the comment analysis method of the present disclosure includes the following steps: searching for multiple user comments from a network database through a crawler module, and storing the multiple user comments in the database; collecting at least one of the multiple user comments through a text detection module, wherein the text detection module includes a character library, and the character library has a plurality of pre-set character strings, and the at least one of the multiple user comments includes at least one of the plurality of character strings; analyzing the at least one of the multiple user comments through a sentiment analysis module; and determining whether the character string has constructive meaning in the at least one of the multiple user comments based on the analysis result of the at least one of the multiple user comments through a determination module.

[0006] Based on the above content, the comment analysis system and comment analysis method disclosed in the present invention can automatically crawl user comments in a network database, and can automatically analyze user comments to determine whether the user comments are constructive.

[0007] In order to make the above contents easier to understand, several embodiments attached with the drawings will be described in detail below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The present disclosure includes accompanying drawings to provide a further understanding of the present disclosure, and the accompanying drawings are incorporated in and constitute a part of this specification. The accompanying drawings illustrate exemplary embodiments of the present disclosure and together with the description serve to explain the principles of the present disclosure.

[0009] Figure 1 is a schematic diagram of a review analysis system according to an embodiment of the present disclosure;

[0010] Figure 2 is a schematic diagram of multiple modules of an embodiment of the present disclosure;

[0011] Figure 3 is a flow chart of a comment analysis method according to an embodiment of the present disclosure;

[0012] Figure 4 is a flowchart of crawling user comments in an embodiment of the present disclosure;

[0013] Figure 5 is a schematic diagram of crawling user comments according to an embodiment of the present disclosure;

[0014] Figure 6 is a schematic diagram of crawling user comments according to another embodiment of the present disclosure;

[0015] Figure 7 is a schematic diagram of building a sentiment analysis model according to an embodiment of the present disclosure;

[0016] Figure 8 is a flow chart of a comment analysis method according to an embodiment of the present disclosure;

[0017] Fig. 9 is a schematic diagram of a construction model and a comment analysis method of an embodiment of the present disclosure;

[0018] Fig.10 is a schematic diagram of a statistical control chart according to an embodiment of the present disclosure.

[0019] Description of Reference Numerals

[0020] 100: Comment analysis system;

[0021] 110: processor;

[0022] 120: Database;

[0023] 130: crawler module;

[0024] 140: text detection module;

[0025] 141: font library;

[0026] 150: analysis module;

[0027] 160: determination module;

[0028] 200: Network database;

[0029] 501~510, 601~610, 801_1~801_P, 902_1~902_S: user comments;

[0030] 701_1~701_N, 901_1~901_R: training data set;

[0031] 702_1~702_M: test data set;

[0032] S310-S340, S410-S450, S810-S880, S910-S960: steps;

[0033] 1001: Control upper limit;

[0034] 1002: Control lower bound;

[0035] 1003:Trend line. DETAILED DESCRIPTION

[0036] Reference will now be made in detail to exemplary embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings. Whenever possible, the same reference numerals are used in the drawings and the description to refer to the same or like parts.

[0037] Throughout the specification and claims of the present disclosure, certain terms are used to refer to specific components. It should be understood by those skilled in the art that electronic device manufacturers may refer to the same component by different names. It is not intended herein to distinguish between components that have the same function but different names. In the following description and claims, words such as "comprise" and "include" are open-ended terms and should be interpreted as "including but not limited to...".

[0038] The term "coupling (coupling) (or connection (connection))" used throughout the entire specification (including the attached claims) of the present disclosure may refer to any direct or indirect connection mode. For example, if the text states that a first device is coupled (or connected) to a second device, it should be interpreted as that the first device can be directly connected to the second device, or that the first device can be indirectly connected to the second device through other devices to be connected or some connection modes. The terms "first (first)", "second (second)" and similar terms mentioned throughout the entire specification (including the attached claims) of this application are only used to name discrete elements or to distinguish different embodiments or scopes. Therefore, the terms should not be regarded as limiting the upper or lower limits of the number of elements and should not be used to limit the order of arrangement of elements. In addition, the elements / components / steps using the same reference numbers in the drawings and embodiments are used as much as possible to represent the same or similar parts. In different embodiments, the same reference numbers or the same terms can be used to refer to the relevant descriptions of the elements / components / steps.

[0039] Figure 1 Schematic diagram of a review analysis system according to an embodiment of the present disclosure. Figure 1 , the review analysis system 100 includes a processor 110 and a database 120. The processor 110 is coupled to the database 120 and is coupled to an external network database 200. In the present embodiment, the network database 200 may be used, for example, to implement one or more network platforms and to sell specific products, such as display panels. The number of network databases 200 is not limited to one. The network database 200 may store user comments edited by multiple users at different times in the process of selling specific products on a web page where their experience after purchasing a specific product is located. In the present embodiment, the review analysis system 100 may be implemented, for example, in the form of a cloud server, but the present disclosure is not limited thereto. The review analysis system 100 may automatically and periodically crawl multiple user comments recorded in the network database 200, and may automatically analyze the multiple user comments to determine whether at least one of the multiple user comments is constructive.

[0040] In this embodiment, the processor 110 may include, for example, a central processing unit (CPU), a graphics processing unit (GPU), or other programmable general-purpose or special-purpose microprocessors, digital signal processors (DSP), programmable controllers, application specific integrated circuits (ASIC), programmable logic devices (PLD), other similar processing devices, or a combination of these devices.

[0041] In the present embodiment, the database 120 includes, for example, any type of fixed or removable random access memory (RAM), read-only memory (ROM), flash memory, hard disk or other circuits or chips with similar functions or a combination of these devices, circuits and chips. The database 120 can be used to store multiple modules, and the multiple modules can be read and executed by the processor 110. The storage device can also be used to store relevant data and information generated during the review analysis process, and / or store relevant data required during the review analysis process. In the present embodiment, the review analysis system 100 may also include a relevant communication interface to connect to the network database 200 via wired or wireless means. In another embodiment of the present disclosure, multiple modules may be built into the processor 110.

[0042] Figure 2 is a schematic diagram of multiple modules of an embodiment of the present disclosure. Figure 3 Flowchart of the comment analysis method of the embodiment of the present disclosure. Figures 1 to 3, the comment analysis system 100 may be pre-built with (stored with) a crawler module 130, a text detection module 140, an analysis module 150 and a determination module 160. In this embodiment, the processor 110 may execute the crawler module 130, the text detection module 140, the analysis module 150 and the determination module 160 to implement the comment analysis method of the following steps S310 to S340. In step S310, the crawler module 130 may search for multiple user comments from the network database 200 and store the multiple user comments in the database 120. In step S320, the text detection module 140 may collect at least one of the multiple user comments, wherein the text detection module 140 includes a character library, and the character library has a plurality of pre-set character strings, and the at least one of the multiple user comments includes at least one of the plurality of character strings. In step S330, the analysis module 150 may analyze the at least one of the multiple user comments. In this embodiment, the analysis module 150 may perform natural language understanding (NLP) on the user comments.

[0043] In step S340, the determination module 160 can determine whether the string has constructive meaning in at least one of the multiple user comments based on the analysis result of at least one of the multiple user comments. Therefore, the comment analysis system 100 of this embodiment can automatically crawl the user comments in the network database 200, and can automatically analyze the user comments to determine whether the user comments have constructive meaning. It is worth noting that the constructive meaning in a specific embodiment means that the user comments include sentences that can reflect the specific defect type for a specific product, so that the user of the comment analysis system 100 can understand the defects of the specific product actually felt by the end user based on the sentences of the specific defect type, and can improve the defects of the specific product. Therefore, for the user of the comment analysis system 100, the user comment has constructive meaning that can improve the specific product, and the comment analysis system 100 can perform data statistics and generate corresponding prompts. The specific implementation methods of the above steps S310 to S340 will be described in detail by the following multiple embodiments.

[0044] Figure 4 1 is a flowchart of crawling user comments in an embodiment of the present disclosure. Figure 1 as well as Figure 4, the crawler module 130 of the processor 110 may perform the following steps S410 to S450. In step S410, the crawler module 130 may obtain the connection information according to the first schedule. The connection information may be, for example, a Uniform Resource Locator (URL) (i.e., a website address). In step S420. The crawler module 130 may store this connection information in the database 120. The first schedule is, for example, in units of months, and may enable the crawler module 130 to crawl and record the connection information corresponding to the multiple user comments month by month. In step S430, the crawler module 130 may select the connection information from the database 120 according to the second schedule. In step S440, the crawler module 130 may connect to the network database 200 according to the selected connection information, and crawl the corresponding multiple user comments in the network database 200. The second schedule is, for example, based on a week, and the crawler module 130 may select the previously stored link information week by week, connect to the network database 200 according to the link information, and crawl the corresponding multiple user comments stored in the network database 200. In step S450, the crawler module 130 may store the multiple user comments in the database 120.

[0045] Specifically, refer to Figure 5 , Figure 5 It is a schematic diagram of crawling user comments according to an embodiment of the present disclosure. The network database 200 may, for example, record user comments 501-510 corresponding to comments 1-10 with different comment contents. The user comments 501-510 also respectively record the comment time t1-t10 when the user writes the comment. The comment time t1-t10 may, for example, include time information of year, month, day, hour, minute, and second, and the comment time t10 is later than the comment time t9, and the comment time t9 is later than the comment time t8. And so on. In this regard, the crawler module 130 may, for example, first obtain the link information of the user comments 501-510 according to a first schedule (for example, monthly), and store the link information of multiple user comments 501-510 in the database 120. Then, the crawler module 130 may select the link information from the database 120 according to a second schedule (for example, weekly), and crawl the user comments in the network database 200 according to the link information. Figure 5As shown, in the previous crawling operation (e.g., the previous week), the crawler module 130 can crawl user comments 501-503, and record the crawling time of the last one of the user comments 501-503 (i.e., record the comment time t3 as the stop time of the previous crawling) for identifying the starting comment of the next crawling. Then, in this crawling operation (e.g., this week), the crawler module 130 can crawl user comments 504-510, and record the crawling time of the last one of the user comments 504-510 (i.e., record the comment time t10 as the stop time of this crawling) for identifying the starting comment of the next crawling.

[0046] refer to Figure 6 , Figure 6 is a schematic diagram of crawling user comments according to another embodiment of the present disclosure. The network database 200 may, for example, record user comments 601-610 corresponding to comments 1-10 with different comment contents. The user comments 601-610 also respectively record the comment time t1-t10 when the user wrote the comment. In this regard, the crawler module 130 may, for example, first obtain the link information of the comments 601-610 according to a first schedule (for example, monthly), and store the link information of multiple user comments 601-610 in the database 120. Then, the crawler module 130 may select the link information from the database 120 according to a second schedule (for example, weekly), and crawl the user comments in the network database 200 according to the link information. Figure 6 As shown, in the previous crawling operation (e.g., the previous week), the crawler module 130 can crawl user comments 601-603 and record the crawling time of the last one of the user comments 601-603 (i.e., record the comment time t3 as the previous crawling stop time) to identify the starting comment for the next crawling. Then, in this crawling operation (e.g., this week), the crawler module 130 can crawl user comments 604-610. However, as Figure 6 As shown, the crawling of user comment 606 may fail, and the crawler module 130 will crawl the user comments 604-610 again in the next schedule. In one embodiment, if the number of crawling failures for the same link information exceeds a preset number, the crawler module 130 will remove the link information from the database 120.

[0047] In addition, the reasons for crawling failure may be, for example, the web page cannot be loaded, the product information cannot be loaded, the comments cannot be sorted by time, the comments that failed to be collected are greater than or equal to a threshold, the comments cannot be limited to the current product, the first page of the non-web page does not have any comments, or the web page cannot switch to the next page, etc., but the present disclosure is not limited to this.

[0048] Figure 7 Schematic diagram of building a sentiment analysis model according to an embodiment of the present disclosure. Figure 1 , Figure 2 as well as Figure 7 In this embodiment, the analysis module 150 may include a sentiment analysis model. The review analysis system 100 may execute the following steps S710 to S740 to build and test the sentiment analysis model. In the model training and construction stage, in step S710, the processor 110 may perform data pre-processing on the training data sets 701_1 to 701_N, where N is a positive integer. To this end, the processor 110 may first annotate the review information in the training data sets 701_1 to 701_N with positive emotions and negative emotions. Positive emotions represent that users have positive emotional feelings towards a specific product. Negative emotions represent that users have negative emotional feelings towards a specific product. Next, the processor 110 may first perform pre-processing operations such as data cleaning, feature processing, and data conversion on the labeled training data sets to eliminate text noise interference before model construction. In this way, it can help simplify the complexity of model construction and accelerate its convergence rate. In step S720, the processor 110 may construct a sentiment analysis model using a Bayesian Classifier based on the pre-processed and labeled training data set. In step S730, the processor 110 may optimize the sentiment analysis model. The processor 110 may evaluate the performance and perform correction optimization on the sentiment analysis model until the model performance of the sentiment analysis model converges to a reliable accuracy.

[0049] Next, in the model prediction stage, at step S710, the processor 110 may perform data pre-processing on the test data sets 702_1 to 702_M, where M is a positive integer. The processor 110 may perform pre-processing operations such as data cleaning, feature processing, and data conversion on the test data sets. At step S720, the processor 110 may input the pre-processed data into the sentiment analysis model. At step S740, the sentiment analysis model may perform sentiment polarity prediction.

[0050] Figure 8 is a flowchart of a comment analysis method according to an embodiment of the present disclosure. Figure 1 , Figure 2 as well as Figure 8 In this embodiment, the processor 110 can execute the crawler module 130, the text detection module 140, the analysis module 150 and the determination module 160 to implement the comment analysis method of the following steps S810 to S880. In this embodiment, the text detection module 140 includes a character library 141. The character library 141 has a plurality of pre-set character strings, and the plurality of character strings are, for example, but not limited to, a plurality of defect code sentences. The plurality of defect code sentences may include, for example, sentences such as screen flashing, bad pixels, bright spots, flowery screens and / or color differences, and the present disclosure is not limited thereto.

[0051] In this embodiment, the processor 110 may execute the crawler module 130 to crawl a plurality of user comments 801_1 to 801_P, where P is a positive integer. In step S810, the processor 110 may execute the text detection module 140 to perform language conversion on the user comments 801_1 to 801_P. For example, the text detection module 140 may convert the user comments 801_1 to 801_P from simplified Chinese to traditional Chinese or English, or the processor 110 may convert the user comments 801_1 to 801_P from traditional Chinese to simplified Chinese, etc., and the present disclosure is not limited thereto. In step S820, the text detection module 140 may search for defective code sentences in the user comments 801_1 to 801_P after language conversion according to a plurality of defective code sentences stored in the character library 141.

[0052] Next, in step S830, the text detection module 140 may determine whether the defective code sentence satisfies the restrictive empirical rule. The text detection module 140 may determine whether the sentence of the user comment after language conversion with the defective code sentence satisfies the restrictive rule to determine whether to perform sentiment analysis on the user comment after language conversion with the defective code sentence. For example, the restrictive empirical rule may be implemented as shown in the following Table 1.

[0053]

[0054] Table 1

[0055] In this regard, if the text detection module 140 determines that the defect code sentence satisfies the restrictive empirical rule, the processor 110 executes step S840 to determine that the current user comment does not have a defect code. On the contrary, if the text detection module 140 determines that the defect code sentence does not satisfy the restrictive empirical rule, the processor 110 executes steps S850 and S860. In step S850, the processor 110 may execute the analysis module 150, such as the above Figure 7 The sentiment analysis module of the embodiment is used to perform complete comment sentiment analysis on the defect code sentences that do not meet the restrictive empirical rules. In addition, in step S860, the processor 110 can execute the analysis module 150, such as the above Figure 7 The sentiment analysis module described in the embodiment performs sentiment analysis on partial comments for defective code sentences that do not meet the restrictive empirical rules. The sentiment analysis on partial comments can be, for example, sentiment analysis on the defective code sentence and the sentences before and after it.

[0056] For example, a user comment may be "I must give a big thumbs up to the merchant's service. The first time the TV was delivered, the screen flickered, and I immediately reported it to the customer service. Because it happened during the Chinese New Year, everyone was on vacation, so they helped to arrange a replacement as soon as the new year began. The after-sales service is great. I have received the TV for almost a month. It is high-definition and the color is very bright. I am very satisfied." The text detection module 140 can compare the character strings in the character library 141 to determine if the above user comment contains the defect code sentence "The first time the TV was delivered, the screen flickered, and I immediately reported it to the customer service." The analysis module 150 can perform partial comment sentiment analysis on the defect code sentence and the sentence before and after it, "I must give a big thumbs up to the merchant's service. The first time the TV was delivered, the screen flickered, and I immediately reported it to the customer service."

[0057] In step S870, the processor 110 may execute the determination module 160 to determine whether the user comment satisfies the defect code rule. In this regard, the determination module 160 may determine whether the defect code rule is satisfied based on the complete comment sentiment analysis results and the partial comment sentiment analysis results. For example, the determination module 160 may perform advanced confirmation of the defect code on the user's evaluation star rating and the sentiment analysis results. In this regard, the determination module 160 may, for example, confirm the defect code rule according to the advanced confirmation rule of thumb summarized in Table 2 below. If the output result of the determination module 160 is "there is a defect code", the determination module 160 may execute step S880 to determine that the current user comment has a defect code. If the output result of the determination module 160 is "there is no defect code", the determination module 160 may execute step S840 to determine that the current user comment does not have a defect code.

[0058] Magnitude Full review sentiment analysis Sentiment analysis of some reviews Output ＞3 front Negative Defective code ≦3 Negative none Defective code ＜3 Negative front Defective code ≦3 front none Defective code ＞3 front front No defect code ＞3 Negative Negative No defect code

[0059] Table 2

[0060] Fig. 9 Schematic diagram of the construction model and comment analysis method of the embodiment of the present disclosure. Figure 1 , Figure 2 as well as Fig. 9In this embodiment, the analysis module 150 may include a trained model. The comment analysis system 100 may perform the following steps S910 to S930 to train the model, wherein the model may be, for example, a neural network model (NN), an artificial intelligence (AI) model, or a chat generative pre-trained transformer (ChatGPT), etc., and the present disclosure is not limited thereto. In this embodiment, in step S910, the processor 110 may process each piece of training data in the training data sets 901_1 to 901_R, where R is a positive integer, to enumerate the potential defect codes therein. In step S920, the processor 110 may train the model using the training data set enumerated with potential defect codes, so that the model can recognize the sentences in the comments and determine whether there are defect code sentences. In step S930, the processor 110 may evaluate the model effect of the trained model, for example, based on the accuracy or related verification indicators to iteratively optimize the model.

[0061] In this embodiment, the processor 110 can execute the crawler module 130, the text detection module 140, the analysis module 150 and the determination module 160 to implement the comment analysis method of the following steps S940 to S960. The crawler module 130 can search for multiple user comments 902_1 to 902_S from the network database 200, where S is a positive integer. In step S940, the text detection module 140 can list potential defect codes for the user comments 902_1 to 902_S to generate an acquisition box string. In step S950, the analysis module 150 can perform model judgment through the aforementioned trained model to determine whether the acquisition box string is a real defect code sentence. In step S960, the determination module 160 can output a judgment result based on the above analysis results.

[0062] Specifically, the text detection module 140 can detect the user comments 902_1~902_S according to the defect code detection table in Table 3 below to list multiple acquisition frame strings. The defect code detection table includes defect code sentences, importance rankings, and synonyms. The text detection module 140 can determine whether there are defect code sentences or their synonyms in the user comments 902_1~902_S, and list the acquisition frame strings according to the importance ranking. The analysis module 150 can analyze the acquisition frame string to determine whether the acquisition frame string is a real defect code sentence, and output the determination result according to the analysis result through the determination module 160.

[0063]

[0064]

[0065] Table 3

[0066] Furthermore, the text detection module 140 can detect the multiple user comments in Table 4 below to obtain a frame string, and the analysis module 150 can determine whether the frame string is a real defective code sentence according to the trained model mentioned above, so as to output a determination result according to the analysis result through the determination module 160. As shown in Table 4 below, the analysis module 150 can accurately determine that the fifth user comment has a real defective code sentence according to the trained model mentioned above.

[0067]

[0068] Table 4

[0069] Fig.10 is a schematic diagram of a statistical control chart according to an embodiment of the present disclosure. Figure 1 , Figure 2 as well as Fig.10 In this embodiment, the determination module 160 can output a determination result according to the above analysis result, and can count multiple user comments corresponding to a specific defect code to generate a result such as Fig.10 Statistical control chart shown. Fig.10 The statistical control chart includes an upper control limit 1001, a lower control limit 1002, and a trend line 1003. Fig.10 As shown, trend line 1003 is used to represent the proportion of a specific product corresponding to a specific defect type, and trend line 1003 may change with statistical results of different years and quarters, and the control upper limit 1001 and the control lower limit 1002 also change dynamically according to the trend.

[0070] In this embodiment, the review analysis system 100 can draw a graph based on the total percentage of negative reviews caused by a specific defect of a specific item and the time series of the reviews. Fig.10 The statistical control chart shown in FIG. 1 automatically monitors whether any data point exceeds the control line (i.e., the upper control limit 1001 and the lower control limit 1002). In this regard, when the trend line 1003 exceeds the upper control limit 1001 in the statistical control chart, or is lower than the lower control limit 1002 in the statistical control chart, the determination module 160 will generate a prompt signal, so as to automatically remind the product manufacturer or product seller of the abnormal trend of a specific defect occurring in the product.

[0071] In addition, in one embodiment, the review analysis system 100 can also provide a conditional query function. For example, a product manufacturer or a product seller can connect to the relevant query interface of the review analysis system 100 and enter any number of keywords for conditions, such as network platform name, product type, review star rating, and defect code, so that the review analysis system 100 can query all user review information that meets the conditions according to the keyword comparison mechanism, and can effectively manage the relevant business of product manufacturing or product sales.

[0072] In summary, the comment analysis system and comment analysis method disclosed in the present invention can automatically crawl user comments in a network database and perform natural language analysis on user comments. The comment analysis system and comment analysis method disclosed in the present invention can perform sentiment analysis and / or model prediction on the results of natural language analysis to effectively identify whether the user comments have defect code sentences. The comment analysis system and comment analysis method disclosed in the present invention can further continuously count the number of defect codes over time and generate a statistical control chart so that corresponding prompt information can be automatically generated according to the statistical control chart.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A review analysis system, It is characterized in that include: A crawler module, coupled to a network database, and configured to search for a plurality of user comments from the network database and store the plurality of user comments in the database; a text detection module, coupled to the crawler module and comprising a character library, wherein the character library has a plurality of pre-set character strings, and the text detection module is used to collect at least one of the plurality of user comments including at least one of the plurality of character strings; an analysis module, coupled to the text detection module and configured to analyze at least one of the plurality of user comments; as well as The determination module is coupled to the analysis module and is used to determine whether the word string has constructive meaning in the at least one of the plurality of user comments according to the analysis result of the at least one of the plurality of user comments.

2. The review analysis system according to claim 1, It is characterized in that The crawler module obtains the connection information according to the first schedule, and the crawler module stores the connection information in the database, The crawler module selects the link information according to a second schedule, crawls a plurality of user comments according to the link information, and stores the plurality of user comments in the database.

3. The review analysis system according to claim 1, It is characterized in that The crawler module crawls the multiple user comments and records the crawling time of the last one of the multiple user comments for identifying the starting comment to be crawled next time.

4. The review analysis system according to claim 1, It is characterized in that The plurality of character strings are a plurality of defective code sentences.

5. The review analysis system according to claim 4, It is characterized in that The text detection module performs language conversion on the plurality of user comments, and searches for at least one of the plurality of language-converted user comments having at least one of the plurality of defective code sentences according to the plurality of defective code sentences.

6. The review analysis system according to claim 5, It is characterized in that The text detection module determines whether at least one of the language-converted sentences of the plurality of user comments having at least one of the plurality of defect code sentences satisfies a restriction expression to determine whether to perform sentiment analysis.

7. The review analysis system according to claim 1, It is characterized in that The analysis module performs a complete comment sentiment analysis and a partial comment sentiment analysis on at least one of the multiple user comments, and the determination module determines whether a defect code rule is satisfied according to the complete comment sentiment analysis result and the partial comment sentiment analysis result.

8. The review analysis system according to claim 1, It is characterized in that The text detection module detects the plurality of user comments according to the defect code detection table to list a plurality of acquisition frame strings. The analysis module analyzes the multiple acquisition frame character strings, and the determination module outputs a determination result according to the analysis result.

9. The review analysis system according to claim 1, It is characterized in that The plurality of user comments correspond to a specific defect code, and the determination module generates a statistical control chart according to the plurality of user comments corresponding to the specific defect code.

10. The review analysis system according to claim 9, It is characterized in that In response to a trend line in the statistical control chart exceeding an upper control limit in the statistical control chart, or falling below a lower control limit in the statistical control chart, the determination module generates a prompt signal.

11. A review analysis method, It is characterized in that include: Searching for multiple user comments from a network database through a crawler module, and storing the multiple user comments in the database; collecting at least one of the plurality of user comments by a text detection module, wherein the text detection module includes a character library, and the character library has a plurality of pre-set character strings, and the at least one of the plurality of user comments includes at least one of the plurality of character strings; analyzing the at least one of the plurality of user comments by a sentiment analysis module; as well as The determination module determines whether the word string has constructive meaning in at least one of the plurality of user comments according to the analysis result of at least one of the plurality of user comments.

12. The review analysis method according to claim 11, It is characterized in that The step of storing the plurality of user comments in the database comprises: obtaining link information according to a first schedule by the crawler module; storing the link information in the database through the crawler module; selecting the link information according to a second schedule by the crawler module; crawling a plurality of user comments according to the link information by the crawler module; and The plurality of user comments are stored in the database through the crawler module.

13. The review analysis method according to claim 11, It is characterized in that The step of storing the plurality of user comments in the database comprises: The crawler module crawls the multiple user comments, and records the crawling time of the last one of the multiple user comments to identify the starting comment to be crawled next time.

14. The review analysis method according to claim 11, It is characterized in that The plurality of character strings are a plurality of defective code sentences.

15. The comment analysis method according to claim 14, It is characterized in that The step of collecting at least one of the plurality of user comments comprises: Performing language conversion on the plurality of user comments by the text detection module; and The text detection module searches for at least one of the language-converted user comments that has at least one of the defective code sentences according to the defective code sentences.

16. The comment analysis method according to claim 15, It is characterized in that The step of collecting at least one of the plurality of user comments further comprises: The text detection module determines whether at least one of the language-converted sentences of the plurality of user comments having at least one of the plurality of defect code sentences satisfies a restriction expression, so as to determine whether to perform sentiment analysis.

17. The review analysis method according to claim 11, It is characterized in that The analysis module performs a complete comment sentiment analysis and a partial comment sentiment analysis on at least one of the multiple user comments, and the step of determining whether the word string has the constructive meaning in at least one of the multiple user comments includes: The determination module determines whether the defect code rule is met based on the complete review sentiment analysis result and the partial review sentiment analysis result.

18. The review analysis method according to claim 11, It is characterized in that Also includes: The text detection module detects the plurality of user comments according to the defect code detection table to list a plurality of acquisition frame strings, and the analysis module analyzes the plurality of acquisition frame strings. The step of determining whether the word string has the constructive meaning in at least one of the plurality of user comments comprises: The analysis module analyzes the multiple acquisition frame character strings, and the determination module outputs a determination result according to the analysis result.

19. The review analysis method according to claim 11, It is characterized in that The plurality of user comments correspond to specific defect codes, and the comment analysis method further comprises: A statistical control chart is generated by the determination module according to the plurality of user comments corresponding to the specific defect code.

20. The review analysis method according to claim 19, It is characterized in that The comment analysis method further includes: In response to the trend line in the statistical control chart exceeding the upper control limit in the statistical control chart, or falling below the lower control limit in the statistical control chart, a prompt signal is generated by the determination module.

Images