TWI567577B - Method of operating a solution searching system and solution searching system - Google Patents

Description

Solution search system operation method and solution search system

The present invention relates to a solution search system, and more particularly to a solution search system that utilizes massive data and data exploration techniques.

In addition to being closely related to R&D technology, the success of the product requires a large number of tests to ensure the stability of the product, especially for high-stability, high-reliability technology products such as industrial instruments, mobile devices, workstations, personal computers. Or products such as servers, etc., the standards for quality control testing are more stringent. When a product is detected, it must be done by copying the problem, collecting and analyzing the relevant information, identifying the cause of the problem, suggesting possible solutions, and verifying the proposed solution... to ensure that the detected problem is Properly resolved, these processes may not only be very time consuming, but may even lead to missed opportunities for the product to enter the market, and implementation must rely on the personal experience and professionalism of the engineer; that is, whether the engineer's experience and professionalism are sufficient will be greatly It affects the time required to propose a solution, and it may also affect whether the proposed solution can completely solve the problem, resulting in the quality of the solution is difficult to grasp. In addition, because personal experience is not easy to pass on, even if the problem to be solved is the same or similar, different engineers may have to repeat the above process to get a solution. This practice is not only inefficient, but also ensures that the engineer can find the most appropriate. s solution.

In addition, for the same type of products, the proportion of the same or similar problems is very high. Although there have been solutions or archived solutions in the past, due to the wide variety of problems, the amount of information involved is quite large, plus Engineers may not be able to describe the problem in a way that is inconsistent, so it is difficult to systematically store it. Therefore, in practice, engineers are still not easy to find relevant solutions, and it is difficult to achieve the purpose of sharing experience with engineers. So how to enable engineers to share their experiences and easily find possible solutions to reduce the time to solve product problems and improve the quality of solutions becomes an important issue.

Embodiments of the present invention provide a solution search system. The solution search system includes a huge database, an associated database, a computing server, a model server, a database server, and a hub server. The computing server is configured to generate a keyword description file according to the standard word comparison table and the text of the problem description file, generate a prediction factor file according to the keyword description file, and generate a model input file according to the prediction factor file. The model server is used to generate a prediction solution category based on the model input file and the data mining prediction model. The database server is configured to read at least one solution from the huge database according to the solution category corresponding to the problem description file. The central server is configured to transmit the problem description file to the computing server when receiving the problem description file, and transmit the model input file generated by the computing server to the model server, and store the problem description file in the associated database. Corresponding to at least one option category, selecting at least one of the at least one option category having the highest weight of the first option category as the solution category, transmitting the solution category to the database server, and each solution according to the associated database The solution corresponds to the weight of the solution category, and at least one solution is sequentially output.

Another embodiment of the present invention provides a method of operating a solution search system. The solution search system includes a computing server, a model server, an associated database, a huge database, a database server, and a hub server. The method includes: when the hub server receives the problem description file, the hub server transmits the problem description file to the operation server, and the operation server compares the text of the standard word comparison table with the problem description file to generate a keyword description file, and the operation The server generates a predictive factor file according to the keyword description file, and the computing server generates a model input file according to the predictive factor file, and the hub server transmits the model input file to the model server, and the model server generates the model input file and the data exploration prediction model according to the model input file. The predictive solution category, when the associated database stores at least one option category corresponding to the problem description file, the hub server selects the first option category with the highest weight among the at least one option category as the solution category, the central server Transmitting the solution category to the database server, the database server reads at least one solution stored in the huge database according to the solution category, and the central server corresponds to each solution according to the associated database Solution category Weights, output sequentially read by a massive library of at least a solution.

1 is a schematic diagram of a solution search system 100 in accordance with an embodiment of the present invention. The solution search system 100 includes a computing server 110, a model server 120, a huge database 130, a database server 140, a hub server 150, and a connected database 160. The database server 140 and the huge database 130 can be used as a database server and a huge database for supporting systems such as Hadoop Distribute File System (HDFS), Hadoop Map/Reduce, and Hive..., or can support other suitable processing. The database system of the data is in compliance with the need for the solution search system 100 to quickly process and store large amounts of data. The associated database 160 (such as MySql, PostgreSql, etc.) can be a general file system, and is used to provide a small amount or temporary data required for the storage of the hub server 150.

When the user wants to use the solution search system 100 to search for a solution, the user can input the problem description file A1 to the solution search system 100. When the hub server 150 receives the problem description file A1, the problem description file A1 can be transmitted to the computing server 110. The problem description file A1 can describe the information related to the product system problem through text description. The content can include the description of the system problem and the phenomenon, the subsystem to which the system problem belongs, and the process of the problem (that is, how to copy the problem), but not Limited to the above information. In an embodiment of the present invention, the problem description file A1 may use a fixed format (such as but not limited to a text file format such as csv, json, xml, etc.) to list information related to system problems, so that the computing server 110 can be more accurate. The interpretation of the problem description file A1 content.

In an embodiment of the present invention, the computing server 110 may generate an attribute description file according to the text of the problem description file A1. The attribute description file may be composed of a plurality of keywords, each of which is composed of a pair of attribute name and attribute value, and is one of the present inventions. In the embodiment, it can be described in the text format of json. When the problem description file A1 uses information related to the system problem in the non-fixed format text bar, the arithmetic server 110 can also use the regular expression to identify the word of the keyword and obtain the value of the keyword. Moreover, the computing server 110 can use the standard word comparison table to compare the text of the problem description file A1 to generate a keyword description file. Table 1 is a partial view of a standard word comparison table according to an embodiment of the present invention. Syntactic vocabulary and vocabulary can be standardized through the standard word comparison table, so that the meaning of the keyword description file can be correctly presented. In addition, in order to avoid the confusion of the semantics of the keyword description file, the values of all keywords can be represented in lowercase.

Table 1         <TABLE border="1" borderColor="#000000" width="_0001"><TBODY><tr><td> "chip set" </td><td> "cs" </td></tr> <tr><td> “chipset” </td><td> “cs” </td></tr><tr><td> “operation system” </td><td> “os” </td ></tr><tr><td> “operation systems” </td><td> “os” </td></tr><tr><td> “system board” </td><td> "sb" </td></tr><tr><td> "mother board" </td><td> "sb" </td></tr><tr><td> "basic input output system </td><td> “bios” </td></tr><tr><td> “memory reference code” </td><td> “mrc” </td></tr><tr ><td> “inter-integrated circuit”, “i2c” </td><td> “i2c” </td></tr></TBODY></TABLE>

In the embodiment of the present invention, the hub server 150 may also update the standard word comparison table according to the input content of the user, because the selection of the keyword may affect the accuracy of the solution search system 100 searching for the solution. To increase the accuracy of the solution search system 100 search solution.

After completing the keyword description file, the computing server 110 may select predictors from the keyword description file to generate a predictor file, and then generate a model input file B1 according to the predictive factor file and the predictive model. For example, the computing server 110 may adjust the predictor profile according to the characteristics of the predictive model (eg, the CBayes model), and delete the digital portion of the predictor file to generate the model input file B1, but different predictive models for the input file There are different requirements for the format, and the present invention is not limited to the above embodiments.

The hub server 150 can transmit the model input file B1 generated by the computing server 110 to the model server 120. The model server 120 can generate the prediction solution category P1 based on the model input file B1 and the data exploration prediction model M1. The predictive solution category P1 can be used to predict the category of the solution to which the problem description file A1 belongs.

In some embodiments of the present invention, the solution category may include a plurality of subcategories, for example, if the prediction solution category P1 is bios.mrc.i2c, where bios indicates that the problem description file A1 may be related to the basic input output system ( Basic Input/Output System, BIOS) related, bios.mrc indicates that the problem description file A1 may be related to the memory reference code in the basic input/output system, and bios.mrc.i2c indicates that the problem description file A1 may be It is related to the Inter-integrated Circuit (I2C) of the memory reference code in the basic input/output system.

Since the prediction solution category P1 generated by the model server 120 is not necessarily correct, when the user has a considerable degree of understanding of the problem description file A1 proposed by the user, the user can also directly correspond to the problem description file A1. The categories are stored in the associated database 160, which stores the user selected option categories in the associated database 160 to increase the chances of finding the correct solution.

In other words, since the different users may encounter the same problem and input the same problem description file when actually searching for the system 100, if the experience and opinions of each user can be integrated, then The accuracy of the solution search system 100 search solution can be further increased. In order to allow the opinions and experiences of different users to be used as a reference for the next search, the solution search system 100 allows the user to enter a self-selected program category corresponding to the problem description file. In addition, considering that each user may have different levels of understanding of the problem, different users may also have different identity weights. Table 2 is a comparison table of the user identity and the weight of the identity of an embodiment of the present invention.

Table 2         <TABLE border="1" borderColor="#000000" width="_0002"><TBODY><tr><td> User Identity</td><td> Weight</td></tr><tr> <td> Project Manager</td><td> 5 </td></tr><tr><td> Technical Manager</td><td> 4 </td></tr><tr>< Td> Senior Engineer</td><td> 3 </td></tr><tr><td> Engineer</td><td> 2 </td></tr><tr><td> Engineer </td><td> 1 </td></tr><tr><td> General User</td><td> 0 </td></tr></TBODY></TABLE >

For example, if the user U1 is a project manager and has an identity weight of 5, when the user U1 inputs a self-selection scheme category corresponding to the problem description file A1, for example, S1, the hub server 150 can select the option category. The correspondence between S1 and the problem description file A1 is stored in the relational database 160, and the weight of the self-selection scheme category S1 corresponding to the problem description file A1 is set according to the identity of the user U1. If no other user inputs the same category before the user U1 inputs the option category S1, the hub server 150 may, in the relational database 160, the weight of the option category S1 corresponding to the problem description file A1. The weight of the user U1 is set to be 5, as shown in Table 2.

After the user U1 inputs the self-selection scheme category S1 corresponding to the problem description file A1, if the user U2 who is the initial engineer also inputs the option category S1 corresponding to the problem description file A1, then the hub server 150 is at this time. Based on the weight of the user U2, which is 1 in Table 2, the weight of the self-selection scheme category S1 corresponding to the problem description file A1 in the relational database 160 is increased, and the weight is increased to 6. In other words, the solution search system 100 can integrate the experience of each user according to the user's identity to increase the accuracy of the solution search system 100 searching for the corresponding solution.

In this way, when selecting the solution category, the hub server 150 first searches the association database 160 for storing at least one option category corresponding to the problem description file A1. If the association type corresponding to the problem description file A1 is not stored in the relational database 160, the hub server 150 selects the prediction solution category P1 generated by the model server 120 as the solution category. In the embodiment of FIG. 1 , since the association database 160 stores the self-selection scheme categories S1 and S2 corresponding to the problem description file A1, the hub server 150 can select the self-selection with the highest weight among the option categories S1 and S2. The scenario category is the solution category C1.

Moreover, in some embodiments, the data mining prediction model M1 can be established by the solution search system 100 based on a plurality of resolved problem description files and data exploration algorithms. The solved problem description file includes the problem cause field, the pairing plan category, and the corresponding solution in addition to the item of the problem description file A1, so the solution search system 100 can be utilized in addition to the option type. The information in the problem description file has been resolved to increase the accuracy of the solution search system 100.

For example, the solution search system 100 may store the correspondence between the pairing scheme category recorded in the resolved problem description file and the corresponding solved problem description file in the associated database 160 as a reference material for subsequent searching. . The pairing plan category may be the solution category that the user originally recorded. However, in some cases, the type of solution recorded by the user may be too large. For example, if the user only uses bios as the solution type of the problem-solving description file, the search solution will be too much and the accuracy will be lost. degree. Therefore, the solution search system 100 can further compare the problem generative field in the solved problem description file with the standard word comparison table to generate a pairing scheme category, thereby improving the accuracy of the pairing scheme category.

In this case, when the association type database corresponding to the problem description file A1 is not stored in the relational database 160, but the pairing scheme category corresponding to the problem description file A1 is stored, the hub server 150 can select the pairing. The scheme category is used as the solution category; when the association database 160 does not store the option category and the pairing scheme category corresponding to the problem description file A1, the hub server 150 selects the prediction solution generated by the model server 120. The scenario category P1 is the solution category.

For example, Table 3 is a correspondence between the problem description file and its self-selection scheme category and pairing scheme category in the association database 160 according to an embodiment of the present invention.

table 3         <TABLE border="1" borderColor="#000000" width="_0003"><TBODY><tr><td> Problem Description File Code</td><td> Option Category</td><td> Pairing Scheme Category</td></tr><tr><td> 01234 </td><td> S1: 6; S2: 3 </td><td> T1 </td></tr><tr> <td> 01235 </td><td> </td><td> T2 </td></tr><tr><td> 01236 </td><td> </td><td> </ Td></tr></TBODY></TABLE>

In Table 3, each problem description file is represented by its problem description file code to facilitate search management. For example, the problem description file code of the problem description file A1 is 01234, and the hub server 150 can search according to the code 01234. Question Description File A1 has a preference category or a pairing scheme category. Since the problem description file A1 corresponds to the resolved problem description file in the solution search system 100, in Table 3, the corresponding pairing plan category T1 can be found according to the problem description file code 01234 of the problem description file A1. Furthermore, according to the problem description file code 01234 of the problem description file A1, the self-selection scheme categories corresponding to the search problem description file A1, that is, S1 and S2, can be found, and the weights are 6 and 3, respectively. In this case, when the hub server 150 selects the solution category of the problem description file A1, the S1 with the higher weight in the option category will be preferentially selected as its solution category.

In other embodiments of the present invention, when the hub server 150 selects the solution category whose problem description file code is 01235, since the option category corresponding to the problem description file code 01235 is not stored in Table 3, There is only a corresponding pairing scheme category T2, so the hub server 150 can select T2 as the solution category for its problem description file. When the hub server 150 selects the solution category whose problem description file code is 01236, since the option category corresponding to the problem description file code 01236 is not stored in Table 3, the file corresponding to the problem description file is not stored. The pairing scheme category of code 01236; that is, the problem description file with the problem description file code 01236 has not been input by the user before, and there is no corresponding solved problem description file in the system for reference, so the hub server 150 will be based on the prediction solution category generated by the model server 120 as its solution category.

After determining the solution category C1, the hub server 150 can transmit the solution category C1 to the database server 140, and the database server 140 can be used to read at least one solution from the huge database 130 according to the solution category C1. Program. In the embodiment of FIG. 1, the solutions D1 ₁ to D1 ₃ corresponding to the solution category C1 are stored in the huge database 130, and the hub server 150 is solved according to each of the association databases 160. The schemes D1 ₁ to D1 ₃ correspond to the weights of the solution category C1, and the solutions D1 ₁ to D1 _{3 are} sequentially output for the user to select. For the user, the greater the weight of the solution indicates that it has a higher probability of solving the problem, so the user can try to solve the problem according to the solution weight from large to small, giving priority to the solution with higher weight. To increase the efficiency of their problem solving.

In some embodiments, the weight of each solution D1 ₁ to D1 ₃ corresponding to the solution category C1 can be automatically set through comparison with the data in the solution search system 100, or can be set by interaction with the user. For example, the solution search system 100 can store the correspondence between the solution category C1 and the respective problem description files corresponding to the solution category C1, that is, the solution search system 100 can record the corresponding to the solution category C1. All resolved problem description files. Since each of the resolved problem description files corresponding to the same solution category C1 may correspond to different solutions, and the solution included in each of the resolved problem description files has an opportunity to be used to solve the problem description file A1 described. The problem, therefore, after the central server selects the appropriate solution category C1, the solution search system 100 can search for each solved problem according to the correspondence between the solution category C1 and its corresponding resolved problem description file. Describe the files and consider the solutions included in each resolved problem description file as a possible solution.

Furthermore, the closer the text content of the resolved problem description file is to the text content of the problem description file A1, the closer the problem described by the two may be, and the more solutions the problem description file contains. The opportunity can solve the problem described in the problem description file A1. Therefore, the solution search system 100 can compare the text content of each solved problem description file with the problem description file A1 being queried, and sort the corresponding problem description files according to the degree of similarity of the comparison results. The weight of the solution.

After the hub server 150 outputs the solutions D1 ₁ to D1 ₃ according to the weight generated by the above manner for the user to select, the user can try to solve the solutions D1 ₁ to D1 ₃ provided by the solution search system 100. problem. In order to allow the experience of different users to be integrated with each other to facilitate the search of the next user, the user can also evaluate the weights of the solutions D1 ₁ to D1 ₃ through the solution search system 100, according to the evaluation results of the user. The hub server 150 can modify the weight of the solution evaluated by the user in the associated database 160 corresponding to the solution category C1.

In other words, the user can give a corresponding weight according to the extent to which the tried solution can solve the problem after actually trying a solution for the problem. For example, Table 4 is a comparison table of the degree and weight of the problem solving according to an embodiment of the present invention.

Table 4         <TABLE border="1" borderColor="#000000" width="_0004"><TBODY><tr><td> Weight</td><td> Degree of problem resolution</td></tr><tr ><td> 0 </td><td> Can't solve the user's problem at all</td></tr><tr><td> 1 </td><td> can provide some clues to solve the user's problem Problem</td></tr><tr><td> 2 </td><td> can provide some clues to solve user problems</td></tr><tr><td> 3 < /td><td> can provide most of the clues to solve user problems</td></tr><tr><td> 4 </td><td> can solve user problems almost</td ></tr><tr><td> 5 </td><td> Solve user problems completely</td></tr></TBODY></TABLE>

In Table 4, the degree to which the solution can be solved according to the tried solution is to completely solve the problem of the user to the problem that the user cannot be solved at all, and can correspond to the weights 5 to 0, respectively. By allowing the user to give a weighted feedback mechanism after trying the solution, the weights of the solutions D1 ₁ to D1 ₃ corresponding to the solution category C1 can be given and accumulated according to the evaluation of different users, so now When a user queries the same system problem, the weights of the solutions D1 ₁ to D1 _{3 are} sequentially output to the user, and the user sequentially tries the solution, so that the user can find the solution to the problem more quickly. The solution, in turn, increases the accuracy of the solution search system 100.

Additionally, the solution search system 100 can also include a web server 170. The user can input the problem description file A1 through the webpage provided by the web server 170, and after receiving the problem description file A1, the web server 170 transmits the problem description file A1 to the hub server 150 and the web interface. The first solutions D1 ₁ to D1 ₃ output by the hub server 150 are displayed on the top.

Through the above-described solution searching system 100 of the embodiment of the present invention, engineers can share their experiences in solving system problems in the past, and can easily find possible solutions to reduce the time for solving product problems and improve the quality of the solution. .

In addition, in order to more effectively utilize the information input by the user, the solution search system of the present invention can also update the prediction model according to the information input by the user, thereby increasing the accuracy of the solution search system.

2 is a schematic diagram of a solution search system 200 according to an embodiment of the present invention. The solution search system 200 and the solution search system 100 can operate according to the same principle, with the difference that the solution search system 200 further includes a modeling server 180. And the problem that the resolved problem description files A2 ₁ to A2 _X received by the hub server 150 reaches a first predetermined number, the standard word comparison table is updated, and/or the user input type of the option category reaches a second predetermined number. The hub server 150 can control the computing server 110 and the modeling server 180 to establish a new data mining prediction model M2.

In FIG. 2, when the solution search system 200 wants to re-establish a new data mining prediction model, the hub server 150 can transmit the resolved problem description files A2 ₁ to A2 _X to the computing server 110, and the computing server file 110 may describe A2 ₁ A2 _X to produce each problem has been described in Item A2 ₁ A2 _X corresponding to the model input file B2 ₁ to B2 _X and solution D2 ₁ from the updated word table and the criterion has been to solve the problem To D2 _X.

The hub server 150 can transmit the solution D2 ₁ to D2 _X corresponding to the problem description files A2 ₁ to A2 _X to the database server 140, so that the database server 140 can be stored in the giant according to the solutions D2 ₁ to D2 _X Volume database 130. At the same time, the hub server 150 can transmit the model input files B2 ₁ to B2 _X and the solution codes D2 ₁ to D2 _X corresponding to the solved problem description files A2 ₁ to A2 _X to the modeling server 180, and model The server 180 can input the file B2 ₁ to B2 _X , the solution codes D2 ₁ to D2 _X, and the associated database description file in the relational database 160 and the selection according to the data exploration algorithm and the plurality of model input files B1 ₁ to B2 _X generated by the computing server 110. The correspondence between the program categories is used to establish a new data mining prediction model M2. In one embodiment of the invention, modeling server 180 may utilize a algorithm suitable for data mining, such as Bayes, CBayes, or SGD, to build a predictive model.

Since the time required to establish the data exploration prediction model M2 is generally much larger than the time required for the model server 120 to convert the data prediction prediction model, the hub server 150 can make the model server only after the data exploration prediction model M2 is established. 120 replaces the data exploration prediction model M1 with the data exploration prediction model M2, so that the user cannot search for the system 200 by using the solution during the establishment of the data prediction prediction model M2. In addition, during the establishment and conversion of the model exploration prediction model M1 by the model server 120, the web server 170 can also output the update progress on the web interface to facilitate the user to know the progress of the model establishment and conversion. . Moreover, the solution search system 200 can also provide a part of the function unrelated to the data mining prediction model for the user to use during the conversion process to increase the convenience of the solution search system 200.

Since the data exploration prediction model M2 is established based on the content input by the previous user, the updated solution search system 200 can integrate the experience of the previous user to improve the accuracy. Through the continuous reconstruction of the data exploration prediction model, the accuracy of the prediction model used in the solution search system of the present invention can be gradually improved and the desired precision can be achieved.

3A and 3B are flow diagrams of a method 300 of operation of the solution search system 100 in accordance with an embodiment of the present invention. The operation method 300 of the solution search system includes steps S310 to S380:

S310: When the hub server 150 receives the problem description file, the hub server 150 transmits the problem description file to the computing server 110, and proceeds to step S320;

S320: The computing server 110 uses the standard word comparison table and the text of the problem description file to generate a keyword description file, and proceeds to step S322;

S322: The computing server 110 generates a predictor file according to the keyword description file, and proceeds to step S324;

S324: The computing server 110 generates a model input file according to the predictive factor file, and proceeds to step S330;

S330. The hub server 150 transmits the model input file to the model server 120, and proceeds to step S340;

S340: The model server 120 generates a prediction solution category according to the model input file and the data exploration prediction model M1, and proceeds to step S350;

S350: If the associated database 160 stores at least one option category corresponding to the problem description file, then proceeds to step S352, otherwise proceeds to step S354;

S352: The hub server 150 selects the first option category with the highest weight among the at least one option category as the solution category, and proceeds to step S360;

S354: If the association database 160 does not store the option category corresponding to the problem description file, but stores the pairing scheme category corresponding to the problem description file, then proceeds to step S356, otherwise proceeds to step S358;

S356: The hub server 150 selects the pairing scheme category as the solution category, and proceeds to step S360;

S358: The hub server 150 selects the prediction solution category generated by the model server 120 as the solution category, and proceeds to step S360;

S360: The hub server 150 transmits the solution category to the database server 140, and proceeds to step S370;

S370: The database server 140 reads at least one solution stored in the huge database 130 according to the solution category, and proceeds to step S380;

S380: The hub server 150 sequentially outputs at least one solution read by the huge database 130 according to the weight of each solution in the association database 160 corresponding to the solution category.

Through the solution searching system 100, 200 and the operation method 300 of the above embodiment of the present invention, the algorithm of the huge amount of database and the data exploration can be utilized to enable the engineers to share their experiences in solving system problems in the past, and the possibility can be easily searched. Solutions to reduce time to resolve product issues and improve solution quality.

In summary, the solution search system and the solution search system operation method of the embodiments of the present invention can utilize a huge database and a data exploration algorithm to assist users in sharing their past experience in solving problems, while using When discovering system problems, it is easy to find possible solutions to reduce the time to resolve product issues. So come. It is possible to avoid the problems in the prior art because the related solutions are not easy to search, and the efficiency and quality of solving the system problems are difficult to control. The above are only the preferred embodiments of the present invention, and all changes and modifications made to the scope of the present invention should be within the scope of the present invention.

100, 200‧‧‧ Solution Search System
110‧‧‧ Computing Server
120‧‧‧Model Server
130‧‧‧ huge database
140‧‧‧Database Server
150‧‧‧Central Server
160‧‧‧Related database
170‧‧‧Web server
180‧‧‧Modeling Server
A1‧‧‧ Problem Description File
B1‧‧‧Model input file
C1‧‧‧Solution category
P1‧‧‧ Forecast Solution Category
S1, S2‧‧‧ Optional category
T1‧‧‧ Matching scheme category
D1 ₁ , D1 ₂ , D1 ₃ ‧ ‧ solutions
A2 ₁ , A2 _X ‧‧‧Resolved problem description file
B2 ₁ , B2 _X ‧‧‧ model input file
D2 ₁ , D2 _X ‧‧‧ solutions
300‧‧‧ method
S310 to S380‧‧‧ steps

FIG. 1 is a schematic diagram of a solution search system according to an embodiment of the present invention. 2 is a schematic diagram of a solution searching system according to another embodiment of the present invention. Figures 3A and 3B are flow diagrams of the method of operation of the solution search system of Figure 1.

S350~S380‧‧‧Steps

Claims (6)

A solution search system, comprising: a huge database; an associated database; an operation server for comparing a text of a standard word with a problem description file to generate a keyword description file, according to The keyword description file generates a predictive factor file, and generates a model input file according to the predictive factor file; a model server is configured to generate a predictive solution category according to the model input file and a first data mining prediction model; a database server for reading at least one solution from the huge database according to the solution category corresponding to the problem description; and a hub server for: when receiving the problem description file, Transmitting the problem description file to the computing server; transmitting the model input file generated by the computing server to the model server; and storing, by the associated database, at least one option corresponding to the problem description file In the category, selecting one of the highest weights in the at least one option category as the first option category a solution category; when the associated database does not store an option category corresponding to the problem description file, but stores a pairing scheme category corresponding to the problem description file, the pairing scheme category is selected as the solution a category; when the associated database does not store the option category corresponding to the problem description file and the pairing scheme category, the prediction solution category is selected as the solution category; Updating the standard word comparison table according to the input content of the user; transmitting the solution category to the database server; and sequentially outputting according to the weight of each solution in the associated database corresponding to the solution category The database server reads the at least one solution from the huge database. The solution search system of claim 1, wherein the hub server is further configured to: set the tried solution according to a problem solving degree of an attempted solution of the at least one solution input by a user The weight of the solution category. The solution search system of claim 1, wherein the hub server is further configured to: when the hub server receives a first predetermined number of resolved problem description files, the standard word comparison table is updated, and/or Controlling the computing server and a modeling server when the user input type of the option is up to a second predetermined number, so that the modeling server generates a file, a data exploration algorithm according to the computing server, and a correspondence between the first option category and the problem description file in the relational database, establishing a second data exploration prediction model; and when the second data exploration prediction model is established, causing the model server to The second data exploration prediction model replaces the first data exploration prediction model. An operation method of a solution search system, the solution search system comprising a computing server, a model server, an associated database, a huge database, a database server and a hub server, the method Include: when the hub server receives a problem description file, the hub server describes the problem file Transmitting to the computing server; the computing server uses a standard word comparison table to compare the text of the problem description file to generate a keyword description file; the computing server generates a prediction factor file according to the keyword description file; The computing server generates a model input file according to the predictive factor file; the hub server transmits the model input file to the model server; the model server generates a model according to the model input file and a first data mining prediction model Predicting a solution category; when the associated database stores at least one option category corresponding to the problem description file, the hub server selects one of the highest weights of the at least one option category as the first option category a solution category; the hub server transmits the solution category to the database server; the database server reads at least one solution stored in the huge database according to the solution category; the hub server According to each solution in the relational database, corresponding to the solution class The weight, sequentially outputting the at least one solution read by the huge database; when a first user inputs the first option category corresponding to the problem description file, the hub server Corresponding relationship between an option category and the problem description file is stored in the associated database, and the weight of the first option category corresponding to the problem description file is set according to the identity of the first user; and After the first user inputs the first option category corresponding to the problem description file, when a second user inputs the first option category corresponding to the problem description file, the hub server is configured according to the second The user's identity weight increases the weight of the first option category corresponding to the problem description file in the associated database. The method of claim 4, further comprising: setting a weight of the tried solution corresponding to the solution category according to a problem solving degree of an attempted solution of the at least one solution input by a user. The method of claim 4, the solution search system further comprising a modeling server, the method further comprising: when the hub server receives a first predetermined number of resolved problem description files, the hub server Updating the standard word comparison table and/or using the input option category for a second predetermined number, the hub server controls the computing server and a modeling server to cause the modeling server to operate according to the operation a file generated by the device, a data exploration algorithm, and a correspondence between the first option category and the problem description file in the associated database, establishing a second data exploration prediction model; and when the second data exploration After the prediction model is established, the hub server causes the model server to replace the first data prediction prediction model with the second data exploration prediction model.