Assessment scale generation method and apparatus
By generating visualized answer results and using a genetic algorithm to select and replace questions, the problem of poor flexibility in traditional paper-based assessment scales is solved, enabling efficient and customized early childhood assessments and providing multi-dimensional assessments and training guidance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HANGZHOU ROBOCT TECH DEV CO LTD
- Filing Date
- 2023-01-09
- Publication Date
- 2026-04-21
AI Technical Summary
Traditional paper-based assessment scales lack flexibility, resulting in inefficient assessments of children's early speech, cognition, hearing, and other functions. They are also difficult to carry, waste resources, and lack customized solutions.
By obtaining the scanning results of the code to be detected, an initial assessment scale is generated, and a visualization result is generated based on the answer results. Genetic algorithms and coding techniques are used to select replacement questions from the behavioral question database to generate a customized final assessment scale, and assessment record information is stored on a remote server.
It improves the efficiency and accuracy of assessments, avoids the phenomenon of children memorizing answers, enables multi-level and targeted assessments, and provides multi-dimensional assessment result displays and training guidance.
Smart Images
Figure CN116166697B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence, and in particular to a method and apparatus for generating assessment scales. Background Technology
[0002] Training children in speech, cognition, and hearing helps identify problems early and allows for targeted training in areas of weakness. Therefore, assessing children in all aspects before training is crucial. However, assessing early childhood speech, cognition, and hearing functions is often challenging. Traditionally, assessors use paper-based scales. Paper-based scales lack flexibility, leading to many children sharing the same scale. Consequently, paper-based scales are inherently inefficient, time-consuming to perform manually, and difficult to carry, resulting in wasted resources.
[0003] Currently, there is no effective solution to the problems of lack of flexibility and customization in the generation of assessment scales in related technologies. Summary of the Invention
[0004] Therefore, it is necessary to provide a method and apparatus for evaluating scale generation to address the aforementioned technical problems.
[0005] Firstly, this application provides a method for generating an assessment scale. The method includes:
[0006] Obtain the code to be detected and the code scanning results for the code to be detected; wherein, the code to be detected includes the interfaces to be collected; obtain the interface class of the code to be detected and the annotation information for the interface class based on the code scanning results, and obtain the initial interface access path based on the annotation information; generate the interface collection results for the interfaces to be collected based on the initial interface access path.
[0007] In one embodiment, after generating an initial assessment scale based on initial behavioral questions, the assessment scale generation method further includes:
[0008] The user's answers to the initial assessment scale are obtained, a visual answer result is generated based on the answer result, and the visual answer result is sent to the display device for display.
[0009] In one embodiment, after generating the initial behavioral questions, the assessment scale generation method further includes:
[0010] Obtain preset filtering rules and user responses to the initial assessment scale; the responses include abnormal responses and non-abnormal responses, where abnormal responses correspond to abnormal behavior questions and non-abnormal responses correspond to non-abnormal behavior questions; retain abnormal behavior questions and obtain replacement behavior questions from the behavior question database according to the filtering rules; generate final behavior questions based on the abnormal behavior questions and replacement behavior questions, and generate the final assessment scale based on the final behavior questions.
[0011] In one embodiment, obtaining replacement behavioral questions from a behavioral question database according to filtering rules includes:
[0012] A preset basic classification standard is obtained, and the segmented behavioral question database is classified according to the basic classification standard to obtain a basic classification result, which includes at least one basic behavioral question category. A preset extended classification standard is obtained, and the basic behavioral question categories are classified according to the extended classification standard to obtain an extended classification result, which includes at least one extended behavioral question category. The basic behavioral question categories are encoded to obtain a basic encoding result, and the basic encoding result is segmented according to the extended behavioral question categories to obtain an extended encoding result. A preset inheritance probability is obtained, and a replacement behavioral question is obtained from the behavioral question data based on the inheritance probability and the extended encoding result.
[0013] In one embodiment, the replacement behavioral question is obtained from the behavioral question data based on the genetic probability and the extended coding result, including:
[0014] Obtain the fitness function, which includes the association between preset question attribute indicators and user data; retrieve the current parent class replacement behavior question from the behavior question database based on the genetic probability and extended encoding results, and retrieve the current child class replacement behavior question from the current parent class replacement behavior question based on the genetic probability; calculate the fitness value of the current parent class replacement behavior question and the fitness value of the current child class replacement behavior question based on the fitness function, compare the fitness values of the current parent class replacement behavior question and the current child class replacement behavior question pairwise, retain the course with the higher fitness value, and generate the next generation of parent class replacement behavior questions; repeat the above steps until the number of iterations reaches the preset maximum number of generations, at which point the iteration ends, and the replacement behavior question is obtained.
[0015] In one embodiment, after comparing the fitness values of the current parent class replacement behavior questions with the fitness values of the current child class replacement behavior questions pairwise and retaining the courses with higher fitness values, the method for generating the next generation of parent class replacement behavior questions further includes:
[0016] Obtain the preset fitness threshold;
[0017] The fitness value of the next generation parent class replacement behavior question is obtained according to the fitness value function. The fitness values of the next generation parent class replacement behavior questions are compared pairwise. If there are two next generation replacement behavior questions whose fitness values are less than the fitness value threshold, the next generation parent class replacement behavior question with the higher fitness value is retained to obtain the deduplication replacement behavior question.
[0018] In one embodiment, after obtaining the reduced-weight replacement questions, the assessment scale generation method further includes:
[0019] Get the preset target number of questions;
[0020] The number of questions with deduplication replacement behavior is detected, and the number of questions with deduplication replacement behavior is compared with the target number of questions to obtain the difference result. If the number of questions with deduplication replacement behavior is less than the target number of questions, the current parent class replacement behavior questions are sorted according to the fitness value to obtain the sorting result. According to the sorting result and the difference result, the current parent class replacement behavior questions are added to the deduplication replacement behavior questions in descending order to obtain the final behavior questions.
[0021] In one embodiment, obtaining the current subclass replacement behavior question from the current parent class replacement behavior question based on the genetic probability includes:
[0022] Obtain the preset crossover probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the crossover probability;
[0023] And / or,
[0024] Obtain the preset mutation probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the mutation probability.
[0025] In one embodiment, the method is applied to a user equipment connected to a remote server; wherein the remote server is also connected to local institutional equipment; the assessment scale generation method further includes:
[0026] The system retrieves assessment record information for the initial assessment scale and sends it to a remote server for storage. Upon receiving a scale retrieval instruction from the local institution's equipment, the remote server sends the assessment record information to the local institution's equipment based on the instruction.
[0027] Secondly, this application also provides an apparatus for generating an assessment scale. The apparatus includes:
[0028] The acquisition module is used to acquire a behavioral question database, category classification rules, and question ratio rules, and to segment the behavioral question database according to the category classification rules to obtain at least one segmented behavioral question database.
[0029] The calculation module is used to acquire user data, filter the corresponding target segmented behavior question database from the segmented behavior question database based on the user data, and randomly select questions from the target segmented behavior question database according to the question ratio rule to generate initial behavior questions;
[0030] The generation module is used to generate an initial assessment scale based on the initial behavioral questions.
[0031] The aforementioned assessment scale generation method and apparatus first acquire a preset behavioral question database, category classification rules, and question ratio rules, and then segment the behavioral question database according to the category classification rules. Second, based on user data, a corresponding target segmented behavioral question database is acquired, and questions are randomly selected from the target segmented behavioral question database according to the question ratio rules to generate initial behavioral questions. Finally, an initial assessment scale is generated based on the initial behavioral questions. This assessment scale generation method solves the problem of difficulty in assessing multiple functions in early childhood using traditional techniques, greatly improving assessment efficiency and avoiding the problem of children memorizing questions. Attached Figure Description
[0032] Figure 1 This is a flowchart illustrating the evaluation scale generation method in one embodiment;
[0033] Figure 2 This is a flowchart illustrating the evaluation scale generation method in another embodiment;
[0034] Figure 3 This is a flowchart illustrating the evaluation scale generation method in another embodiment;
[0035] Figure 4 This is a flowchart illustrating the assessment scale generation method in a preferred embodiment;
[0036] Figure 5 This is a structural block diagram of an assessment scale generation device in one embodiment. Detailed Implementation
[0037] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0038] In one embodiment, such as Figure 1 As shown, a method for generating an assessment scale is provided. Figure 1 This is a flowchart of an assessment scale generation method according to an embodiment of this application, the process including the following steps:
[0039] Step S110: Obtain the behavior question database, category division rules, and question ratio rules. Divide the behavior question database into segments according to the category division rules to obtain at least one segmented behavior question database.
[0040] The behavioral question database in this application uses a child disability screening and assessment scale, which is an internationally recognized assessment system. The pre-defined categorization rules in this application may include dividing the question database according to age groups or according to user software usage. For example, when dividing the behavioral question database by age groups, the questions in the database differ for different age groups, and the difficulty of the questions increases with age. Dividing the behavioral question database according to this categorization rule results in multiple segmented behavioral question databases.
[0041] Step S120: Obtain user data; based on the user data, filter the corresponding target segmented behavior question database from the segmented behavior question database; randomly select questions from the target segmented behavior question database according to the question ratio rule to generate initial behavior questions.
[0042] Specifically, the corresponding target segmented behavioral question database is obtained by filtering from the segmented behavioral question database based on user data. The user data may include information such as user age and user software usage records. For example, when the user data is user age, the corresponding target segmented behavioral question database can be filtered according to the preset division of the behavioral question database into different age groups, and then initial behavioral questions are randomly obtained from the target segmented behavioral question database according to the preset question ratio rules.
[0043] Step S130: Generate an initial assessment scale based on the initial behavioral questions.
[0044] Through steps S110 to S130, firstly, a preset behavioral question database, category classification rules, and question ratio rules are obtained. The category classification rules include dividing the behavioral question database according to different age groups of users, resulting in multiple segmented behavioral question databases. Furthermore, the questions and difficulty levels in the segmented behavioral question databases differ across age groups; as age increases, the difficulty of the questions in the segmented behavioral question database also increases. Secondly, after obtaining user data, the corresponding target segmented behavioral question database is obtained by filtering from the segmented behavioral question database based on the user's age information in the user data. Questions are randomly selected from the target segmented behavioral question database according to the preset question ratio rules to generate initial behavioral questions. Finally, an initial assessment scale is generated based on the initial behavioral questions. In addition to dividing the database by age group, this behavioral question database can also be categorized according to preset basic classification criteria. In this application, it can be divided into three main categories: hearing, language, and cognition. These basic categories can be further subdivided according to preset extended classification criteria. For example, the language category can include extended categories such as language development delay and articulation disorders. The number of questions varies across different categories. Therefore, compared to traditional assessment scales, the assessment scale in this application provides multi-faceted and multi-level evaluation, effectively addressing the issue of children memorizing answers. Furthermore, by completely eliminating paper-based assessment scales, it improves the efficiency and convenience of assessment. Moreover, since this application uses user data to obtain the corresponding target-segmented behavioral question database, it enhances the targeting and accuracy of the assessment, achieving efficient assessment for different users.
[0045] In one embodiment, Figure 2 This is a flowchart illustrating the assessment scale generation method in another embodiment. The assessment scale generation method further includes:
[0046] Step S240: Obtain the user's answer results for the initial assessment scale based on the initial assessment scale, generate a visual answer result based on the answer result, and send the visual answer result to the display device for display.
[0047] Specifically, the assessment scale generation method provided in this application can also generate visualized response results based on the answering results, which can be displayed in multiple dimensions through radar charts and other formats. These visualized response results include both abnormal and normal answering results, and can further display the user's specific assessment performance in different aspects, summarize the overall assessment results, and highlight issues requiring attention. These visualized response results can be generated automatically, allowing users and professionals to clearly observe and analyze the responses. Based on the assessment results, professionals can provide subsequent behavioral training guidance.
[0048] In one embodiment, Figure 3 This is a flowchart illustrating the assessment scale generation method in another embodiment. The assessment scale generation method further includes:
[0049] Step S350: Obtain the preset filtering rules and the user's answers to the initial assessment scale; the answer results include abnormal answer results and non-abnormal answer results, wherein abnormal answer results correspond to abnormal behavior questions and non-abnormal answer results correspond to non-abnormal behavior questions; retain the abnormal behavior questions, and obtain replacement behavior questions from the behavior question database according to the filtering rules; generate the final behavior questions based on the abnormal behavior questions and the replacement behavior questions, and generate the final assessment scale based on the final behavior questions.
[0050] Specifically, after obtaining user answers, the system retains the abnormal behavior questions corresponding to the abnormal results, while replacing the questions without abnormal behavior. When replacing questions without abnormal behavior, replacement questions are retrieved from the behavior question database according to preset filtering rules. Different proportions are used for different question types within different categories when retrieving replacement questions. If all questions in a major category of the basic classification results are without abnormalities, the difficulty of the replacement behavior questions is increased according to the preset question proportion rules when generating the assessment scale again. If questions in an extended category of the basic classification results are without abnormalities, the extended category results without abnormalities are redistributed with increased difficulty according to the preset question proportion rules. It can be seen that this application completely avoids the situation of children memorizing questions. Correctly answered questions are deleted, and more difficult questions of the same type are added according to a preset proportion. After multiple comparative analyses of assessments, a professional assessment report is generated. Timely increases in question difficulty can more efficiently obtain accurate assessment results, ensuring the sensitivity and accuracy of the assessment.
[0051] In one embodiment, the above-described method for generating assessment scales further includes:
[0052] Obtain the preset basic classification criteria, classify the segmented behavioral question database according to the basic classification criteria, and obtain the basic classification results, which include at least one basic behavioral question category.
[0053] Obtain a preset extended classification standard, classify the basic behavioral question categories according to the extended classification standard, and obtain extended classification results, which include at least one extended behavioral question category;
[0054] The basic behavioral question categories are coded to obtain basic coding results. The basic coding results are then segmented according to the extended behavioral question categories to obtain extended coding results.
[0055] Obtain the preset genetic probability, and then obtain the replacement behavioral question from the behavioral question data based on the genetic probability and the extended encoding result.
[0056] Specifically, in this application, based on the age-group-based database, the behavioral question database can also be categorized according to preset basic classification standards. In this application, it can be divided into three main categories: hearing, language, and cognition. Furthermore, the basic classification results can be further subdivided according to preset extended classification standards. For example, the language category can include extended classification results such as language development delay and articulation disorders. The number of questions varies across different categories. Multi-stage categorization of the behavioral question database and setting different numbers of questions for different types can yield more comprehensive assessment results, clearly showcasing the user's assessment results in various aspects. Further, the hearing assessment question bank references "Standards and Methods for Assessing Hearing Ability in Children with Hearing Impairment - Standards and Methods for Assessing Language Ability" and "Speech Therapy." Hearing assessment primarily evaluates whether the user has hearing impairment and language abnormalities. Hearing-impaired users have impairments in their articulation organs, often unable to obtain their true sound level or adjust their own vocal feedback; therefore, some language assessments are also conducted for children with hearing impairments. The language assessment question bank references the SS scale to assess users' articulation disorders and language development delays. The assessment results statistically indicate that users with language disorders may have suspected articulation organs and disorders, manifested in respiratory function (inability to control airflow leading to incorrect pronunciation), weakened oral muscles and tongue function leading to speech disorders, including lipization, unclear vowel pronunciation, language delay, and misunderstanding of words and sentences. The cognitive assessment question bank references the Child Development Assessment Scale. Cognitive assessment primarily evaluates whether users have problems with self-care abilities, language abilities, sensory and cognitive abilities, and fine and gross motor skills. This demonstrates that the preliminary investigation of this application involved extensive research into international standard assessment methods, systematically combining numerous assessment standards from different perspectives, resulting in a comprehensive assessment that utilizes multiple rounds of evaluation to obtain accurate results.
[0057] In one embodiment, the above-mentioned assessment scale generation method further includes:
[0058] Obtain the fitness function, which includes the association between preset question attribute indicators and user data; retrieve the current parent class replacement behavior question from the behavior question database based on the genetic probability and extended encoding results, and retrieve the current child class replacement behavior question from the current parent class replacement behavior question based on the genetic probability; calculate the fitness value of the current parent class replacement behavior question and the fitness value of the current child class replacement behavior question based on the fitness function, compare the fitness values of the current parent class replacement behavior question and the current child class replacement behavior question pairwise, retain the course with the higher fitness value, and generate the next generation of parent class replacement behavior questions; repeat the above steps until the number of iterations reaches the preset maximum number of generations, at which point the iteration ends, and the replacement behavior question is obtained.
[0059] Specifically, a preset fitness function is obtained, which is constructed as follows:
[0060]
[0061] Among them, f i ω represents the absolute value of the error between the i-th attribute indicator and the user's requirement. iLet represent the weight of the importance of the i-th attribute indicator to the course, j be the number of attribute indicators participating in the evaluation, and f be the sum of the absolute values of the errors between all indicators and the user's requirements. Each question in the behavioral question database has multiple attribute indicators, including question name, question options, question answer, level information, difficulty level, etc. The smaller the fitness value, the more suitable the question is for the user's needs. It can be seen that the fitness function focuses on calculating whether the behavioral question is close to the user's needs. Specifying multiple attribute indicators for each behavioral question is beneficial for comprehensively calculating each question and obtaining the most suitable result. It also lays the foundation for iteratively replacing behavioral questions based on the fitness value of the question in subsequent steps. The genetic probability in this application includes crossover probability and mutation probability. In the encoding result, the current parent class replacement behavioral question is obtained from the behavioral question database according to the genetic probability, and the current child class replacement behavioral question is formed through crossover, mutation, and other operations. Then, the fitness values of the current parent class replacement behavioral question and the current child class replacement behavioral question are calculated according to the preset fitness function, and the two are compared pairwise. The course with the higher fitness value is retained to generate the next generation of parent class replacement behavioral questions. Repeat the above steps until the number of iterations equals the preset maximum generation, at which point the iteration ends, yielding the replacement behavior items. Furthermore, since the number of items varies across different base classifications, the maximum generation also differs; the larger the number of items, the larger the maximum generation. Therefore, this application proposes converting a large database of behavior items into codes, performing crossover and mutation calculations on the codes to iterate the generated replacement behavior items. This provides targeted, customized items for different users. The assessment scales performed by users at each stage are automatically updated and generated according to the above steps, and the assessment results are also displayed in multi-dimensional charts generated by the system. Assessment items and methods do not repeat, preventing a decrease in user sensitivity to the assessment scale and resulting in more accurate assessment results.
[0062] In one embodiment, the above-described method for generating assessment scales further includes:
[0063] Obtain the preset fitness threshold;
[0064] The fitness value of the next generation parent class replacement behavior question is obtained based on the fitness value function. The fitness values of the next generation parent class replacement behavior questions are compared pairwise. If the difference between the fitness values of two next generation replacement behavior questions is less than the fitness value threshold, the next generation parent class replacement behavior question with the higher fitness value is retained, and the deduplication replacement behavior question is obtained.
[0065] Specifically, when generating replacement behavior questions, deduplication is performed. The fitness values of the next-generation parent class replacement behavior questions are compared pairwise. If the difference in fitness values between two questions is less than a preset fitness threshold, the question with the higher fitness value is retained, thus obtaining deduplicated replacement behavior questions. It can be seen that the deduplication step enables the algorithm of this application to complete iterations more smoothly and converge quickly. Furthermore, by promptly removing unnecessary questions, it reduces unnecessary computational burden and significantly improves computational efficiency.
[0066] In one embodiment, the method further includes:
[0067] Get the preset target number of questions;
[0068] The number of questions with deduplication replacement behavior is detected, and the number of questions with deduplication replacement behavior is compared with the target number of questions to obtain the difference result. If the number of questions with deduplication replacement behavior is less than the target number of questions, the current parent class replacement behavior questions are sorted according to the fitness value to obtain the sorting result.
[0069] Based on the sorting results and the difference results, the current parent class replacement behavior questions are added to the deduplication replacement behavior questions in descending order to obtain the final behavior questions.
[0070] Specifically, since the number of different categories of items in the assessment scale is fixed, and the total number of items in the scale is also fixed, the assessment scale needs to be supplemented after deduplication. This supplementation can be achieved by obtaining items from the current parent category replacement behavior items. The current parent category replacement items are sorted according to their fitness values. After sorting, based on the sorting results and the difference results, the current parent category replacement behavior items are added to the deduplication replacement behavior items in ascending order to obtain the final behavior items. Because the number of replacement behavior items changes after deduplication, new items need to be added promptly to maintain a constant number of items. Supplementing items allows for obtaining as much user information as possible within a limited number of assessments, thereby generating more accurate assessment results, avoiding wasted assessment opportunities, and ensuring that the proportion of each type of item in the generated items remains constant, efficiently completing the user assessment.
[0071] In one embodiment, the method further includes:
[0072] The current parent class replacement behavior question is used to retrieve the current child class replacement behavior question, including:
[0073] Obtain the preset crossover probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the crossover probability;
[0074] And / or,
[0075] Obtain the preset mutation probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the mutation probability.
[0076] Specifically, the crossover operation includes the crossover probability mentioned above. The crossover operation directly swaps the two encoded items bit by bit according to a certain crossover probability. Considering convergence speed, global search, and incorporating small-scale competition and selection operations, each current parent class replacement behavior item is generated with only one individual, ensuring the offspring population size remains unchanged. The selected current parent class replacement behavior items are randomly assigned, and the same number of current subclass courses are generated through crossover, resulting in current subclass replacement behavior items. And / or, the mutation operation includes the mutation probability mentioned above. The mutation operation inverts the encoding of the behavior training course according to a certain mutation probability. The expanded encoding result represents a certain number of expanded classification results. Mutation cannot change the size of each expanded encoding result, and the mutation operation can only be performed on an expanded encoding result basis, selecting the item to be replaced from the same type of item according to a certain mutation probability. Therefore, it can be seen that after generating the initial evaluation scale, the initial evaluation scale is transformed according to the preset classification criteria and genetic probability to obtain the final evaluation scale. Furthermore, this genetic probability includes crossover probability and mutation probability. After multiple iterations, the final assessment scale has digital statistics and a customized, phased assessment arrangement, forming a closed loop of "assessment-report-training-reassessment-retraining". Compared with the generation of traditional behavioral assessment scales, the customized assessment scale generated in this application can achieve better assessment results.
[0077] In one embodiment, the method further includes:
[0078] The system retrieves assessment record information for the initial assessment scale and sends it to a remote server for storage. Upon receiving a scale retrieval instruction from the local institution's equipment, the remote server sends the assessment record information to the local institution's equipment based on the instruction.
[0079] Specifically, the aforementioned assessment record information includes the assessment results of all users on each user terminal and the corresponding assessment scales, which are uniformly stored in a remote server. The aforementioned scale retrieval command is a retrieval command issued by a specific user terminal. When the remote server receives this command, it retrieves the corresponding user's assessment results and the corresponding assessment scales from the assessment record information and sends the user's assessment results and assessment scales to the corresponding user terminal, which may include local institutional devices. In summary, the cloud data storage and intelligent analysis functions of this application can synchronize assessment and training data from other terminals, as well as user training data at institutions or at home, providing data support for intelligent assessment. This application integrates a large number of assessment results, which not only helps users observe the comprehensive assessment results more clearly and more intuitively identify items that need further training, but also integrates data from multiple users, further optimizing the assessment scale generation method in this application.
[0080] This embodiment also provides a specific embodiment of an assessment scale generation method, such as... Figure 4 As shown, Figure 4 This is a schematic diagram of the process for generating an assessment scale in a preferred embodiment.
[0081] Questions are extracted from the behavior question database based on the user's age information to obtain initial behavior questions. Specifically, the behavior question database is divided according to a preset classification rule, which can be based on age groups, resulting in multiple segmented behavior question databases. The corresponding database is then selected from the segmented behavior question databases based on the user's input age information, and questions are randomly selected from the target segmented behavior question database according to a preset question ratio rule to obtain the initial behavior questions.
[0082] The segmented behavioral question database is categorized according to a pre-defined classification standard. A basic classification standard is obtained, resulting in a basic classification result. This basic classification standard includes division into three main categories: listening, language, and cognition. Further, based on the basic classification result, more detailed divisions are made according to pre-defined extended classification results. For example, the language basic classification result is further divided into extended classification results for language development delay and articulation disorders. The number of questions varies in different categories. After classification, the courses for each basic classification result are encoded. This encoding includes, but is not limited to, binary encoding, and the encoding is divided into several sub-segments according to the extended classification standard, with each sub-segment having an independent code.
[0083] The process involves retaining the abnormal behavior questions corresponding to abnormal answers from the user's responses and replacing the non-abnormal behavior questions with the non-abnormal behavior questions corresponding to non-abnormal answers, resulting in replacement behavior questions. A final evaluation scale is then derived based on these abnormal behavior questions and the replacement behavior questions. Specifically, replacing non-abnormal behavior questions with replacement behavior questions requires a pre-defined genetic probability and fitness function. First, a fitness function is introduced. Each behavior question includes at least one question attribute indicator, such as type, options, difficulty level, and level information. Based on the pre-defined fitness function, the fitness level of each behavior question is calculated according to the question attribute indicators; the lower the fitness value, the stronger the fitness of the question. Then, based on the genetic probability and fitness function, and using this extended encoding result, the current parent-class replacement behavior question is retrieved from the target behavior question database, and the current child-class replacement behavior question is retrieved from the current parent-class replacement behavior question based on the genetic probability. The fitness values of the current parent class replacement behavior question and the current child class replacement behavior question are calculated based on the fitness function. The fitness values of the current parent class replacement behavior question and the current child class replacement behavior question are compared pairwise, and the course with the higher fitness value is retained to generate the next generation of parent class replacement behavior questions. The next generation of parent class replacement behavior questions is then de-duplicated and supplemented. The above steps are repeated until the preset maximum number of iterations is reached, at which point the iteration ends, and the final behavior question is obtained.
[0084] The questions and assessment results are transmitted to a remote server for storage. The remote server is connected to the local facility's equipment and has data storage and intelligent analysis capabilities. Upon receiving a scale retrieval instruction from the local facility, the remote server sends the stored records to the local facility.
[0085] As can be seen from this, this application automatically generates assessment scales based on dynamic question generation algorithms, uses intelligent rehabilitation algorithms for searching and optimization, employs encoding techniques to represent various complex question bank structures, introduces fitness function, calculates the average fitness function of the question group, and uses genetic operations and survival of the fittest through encoding representation to determine the search results, and iteratively calculates to form the optimal assessment questionnaire.
[0086] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0087] Based on the same inventive concept, this application also provides an apparatus for generating assessment scales as described above. The solution provided by this apparatus is similar to the solution described in the above method; therefore, the specific limitations of one or more assessment scale generation apparatus embodiments provided below can be found in the limitations of the assessment scale generation method described above, and will not be repeated here.
[0088] In one embodiment, such as Figure 5 As shown, an assessment scale generation device is provided, including: an acquisition module 51, a calculation module 52 and a generation module 53.
[0089] The acquisition module 51 is used to acquire a behavioral question database, a category division rule, and a question ratio rule, and to segment the behavioral question database according to the category division rule to obtain at least one segmented behavioral question database.
[0090] Calculation module 52 is used to acquire user data, filter the corresponding target segmented behavior question database from the segmented behavior question database based on the user data, randomly select questions from the target segmented behavior question database according to the question ratio rule, and generate initial behavior questions.
[0091] The generation module 53 is used to generate an initial assessment scale based on the initial behavioral questions.
[0092] Specifically, the acquisition module 51 is connected to the calculation module 52. The acquisition module 51 acquires the behavioral question database, category classification rules, and question ratio rules. Further, the acquisition module 51 segments the behavioral question database according to the category classification rules to obtain a segmented behavioral question database. The acquisition module 51 sends the segmented behavioral question database and the question ratio rules to the calculation module 52. The calculation module 52 filters the segmented behavioral question database based on user data to obtain a target segmented behavioral question database, thus obtaining initial behavioral questions. The calculation module 52 sends the initial behavioral questions to the generation module 53, which generates an initial assessment scale based on the initial behavioral questions.
[0093] The aforementioned assessment scale generation device differs from traditional assessment scale generation methods, completely eliminating the need for children to memorize questions in traditional assessment scales. It automatically generates assessment scales, reflecting both the professionalism and scientific nature of the assessment, while ensuring the sensitivity and accuracy of the assessment. Furthermore, after the assessment is completed, it facilitates professional personnel to provide corresponding training and guidance based on the assessment results, thus providing users with a digital assessment system.
[0094] Each module in the aforementioned behavioral assessment scale generation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0095] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0096] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for generating an assessment scale, characterized in that, The method includes: Obtain a behavioral question database, category classification rules, and question ratio rules. Segment the behavioral question database according to the category classification rules to obtain at least one segmented behavioral question database. Acquire user data, and based on the user data, filter the corresponding target segmented behavior question database from the segmented behavior question database. Randomly select questions from the target segmented behavior question database according to the question ratio rule to generate initial behavior questions. An initial assessment scale is generated based on the initial behavioral questions. After generating the initial behavioral question, the method further includes: Obtain the preset filtering rules and the user's answers to the initial assessment scale; The answer results include abnormal answer results and non-abnormal answer results, wherein the abnormal answer results correspond to questions with abnormal behavior, and the non-abnormal answer results correspond to questions with no abnormal behavior; The abnormal behavior questions are retained, and replacement behavior questions are obtained from the behavior question database according to the filtering rules. Based on the abnormal behavior questions and the alternative behavior questions, final behavior questions are generated, and a final assessment scale is generated based on the final behavior questions. The step of obtaining replacement behavioral questions from the behavioral question database according to the filtering rules includes: Obtain a preset basic classification standard, classify the segmented behavioral question database according to the basic classification standard, and obtain a basic classification result, wherein the basic classification result includes at least one basic behavioral question category; Obtain a preset extended classification standard, classify the basic behavior question categories according to the extended classification standard, and obtain an extended classification result, wherein the extended classification result includes at least one extended behavior question category; The basic behavior question categories are encoded to obtain basic encoding results. The basic encoding results are then segmented according to the extended behavior question categories to obtain extended encoding results. Obtain a preset genetic probability, and derive the replacement behavioral question from the behavioral question data based on the genetic probability and the extended encoding result; The step of obtaining the replacement behavioral question from the behavioral question data based on the genetic probability and the extended coding result includes: Obtain the fitness function, which includes the association between preset question attribute indicators and user data; Based on the genetic probability and the extended encoding result, the current parent class replacement behavior question is obtained from the behavior question database, and the current child class replacement behavior question is obtained from the current parent class replacement behavior question based on the genetic probability; The fitness value of the current parent class replacement behavior question and the fitness value of the current child class replacement behavior question are calculated according to the fitness value function. The fitness values of the current parent class replacement behavior question and the current child class replacement behavior question are compared pairwise. The course with the higher fitness value is retained to generate the next generation of parent class replacement behavior questions. The above steps are repeated until the number of iterations reaches the preset maximum number of generations, and the iteration ends to obtain the replacement behavior question.
2. The method according to claim 1, characterized in that, After generating the initial assessment scale based on the initial behavioral questions, the method further includes: The user's answers to the initial assessment scale are obtained based on the initial assessment scale. A visual answer result is generated based on the answer result and sent to a display device for display.
3. The method according to claim 1, characterized in that, After comparing the fitness values of the current parent class replacement behavior question with the fitness values of the current child class replacement behavior question pairwise, retaining the courses with higher fitness values, and generating the next generation of parent class replacement behavior questions, the method further includes: Obtain the preset fitness threshold; The fitness value of the next generation parent class replacement behavior question is obtained according to the fitness value function. The fitness values of the next generation parent class replacement behavior questions are compared pairwise. If there are two fitness values of the next generation parent class replacement behavior questions with a difference less than the fitness value threshold, the next generation parent class replacement behavior question with the higher fitness value is retained to obtain the deduplication replacement behavior question.
4. The method according to claim 3, characterized in that, After obtaining the deduplication replacement behavior question, the method further includes: Get the preset target number of questions; The number of questions with the deduplication replacement behavior is detected, and the number of questions with the deduplication replacement behavior is compared with the target value of the number of questions to obtain the difference result. If the number of questions with the deduplication replacement behavior is less than the target value of the number of questions, the current parent class replacement behavior questions are sorted according to the fitness value to obtain the sorting result. Based on the sorting result and the difference result, the current parent class replacement behavior question is added to the deduplication replacement behavior question in descending order to obtain the final behavior question.
5. The method according to claim 1, characterized in that, The step of obtaining the current subclass replacement behavior question from the current parent class replacement behavior question based on the genetic probability includes: Obtain the preset crossover probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the crossover probability; And / or, Obtain the preset mutation probability, and based on the extended encoding result, obtain the current subclass replacement behavior question from the current parent class replacement behavior question according to the mutation probability.
6. The method according to any one of claims 1 to 5, characterized in that, The method is applied to user equipment, which is connected to a remote server; wherein the remote server is also connected to local facility equipment; the method further includes: The assessment record information for the initial assessment scale is obtained and sent to a remote server for storage; wherein, when the remote server receives a scale acquisition instruction from the local institution device, it sends the assessment record information to the local institution device based on the scale acquisition instruction.
7. An assessment scale generation device, characterized in that, The device includes: The acquisition module is used to acquire a behavioral question database, category classification rules, and question ratio rules, and to segment the behavioral question database according to the category classification rules to obtain at least one segmented behavioral question database. The calculation module is used to acquire user data, filter the corresponding target segmented behavior question database from the segmented behavior question database based on the user data, and randomly select questions from the target segmented behavior question database according to the question ratio rule to generate initial behavior questions; The generation module is used to generate an initial assessment scale based on the initial behavioral questions; After generating the initial behavior question, the process also includes: Obtain the preset filtering rules and the user's answers to the initial assessment scale; The answer results include abnormal answer results and non-abnormal answer results, wherein the abnormal answer results correspond to questions with abnormal behavior, and the non-abnormal answer results correspond to questions with no abnormal behavior; The abnormal behavior questions are retained, and replacement behavior questions are obtained from the behavior question database according to the filtering rules. Based on the abnormal behavior questions and the alternative behavior questions, final behavior questions are generated, and a final assessment scale is generated based on the final behavior questions. The step of obtaining replacement behavioral questions from the behavioral question database according to the filtering rules includes: Obtain a preset basic classification standard, classify the segmented behavioral question database according to the basic classification standard, and obtain a basic classification result, wherein the basic classification result includes at least one basic behavioral question category; Obtain a preset extended classification standard, classify the basic behavior question categories according to the extended classification standard, and obtain an extended classification result, wherein the extended classification result includes at least one extended behavior question category; The basic behavior question categories are encoded to obtain basic encoding results. The basic encoding results are then segmented according to the extended behavior question categories to obtain extended encoding results. Obtain a preset genetic probability, and derive the replacement behavioral question from the behavioral question data based on the genetic probability and the extended encoding result; The step of obtaining the replacement behavioral question from the behavioral question data based on the genetic probability and the extended coding result includes: Obtain the fitness function, which includes the association between preset question attribute indicators and user data; Based on the genetic probability and the extended encoding result, the current parent class replacement behavior question is obtained from the behavior question database, and the current child class replacement behavior question is obtained from the current parent class replacement behavior question based on the genetic probability; The fitness value of the current parent class replacement behavior question and the fitness value of the current child class replacement behavior question are calculated according to the fitness value function. The fitness values of the current parent class replacement behavior question and the current child class replacement behavior question are compared pairwise. The course with the higher fitness value is retained to generate the next generation of parent class replacement behavior questions. The above steps are repeated until the number of iterations reaches the preset maximum number of generations, and the iteration ends to obtain the replacement behavior question.
Citation Information
Patent Citations
Psychological evaluation system scale random answering method and system
CN113010557A