College student personalized growth path planning method based on AI big data

Abstract

The invention provides a university student personalized growth path planning method based on AI big data, and belongs to the technical field of education informatization and artificial intelligence. The method comprises the following steps: constructing a full-link multi-source data system running through enrollment, cultivation and employment, and integrating and associating biogenic features, learning processes and post demand data to form a full-cycle unified data basis; constructing a capability association model based on the system, and establishing a mapping relation between student capabilities and post demands; a personalized growth path is generated based on the model, and the path contains a development direction recommendation of capability and demand matching and can be dynamically adjusted according to a preset condition. According to the method, a precise data source is provided through full-link data association, man-post quantitative matching is realized through a capability model, and the college talent training precision and the student employment competitiveness are effectively improved by dynamically adjusting and adapting growth and industry changes.

Description

Technical Field

[0001] This invention relates to the fields of educational informatization and artificial intelligence technology, and in particular to a method for planning personalized growth paths for college students based on AI big data. Background Technology

[0002] The cultivation of engineering talents in universities is a key link in supporting the country's high-level scientific and technological self-reliance and building a modern industrial system. Its quality directly determines the development momentum of emerging industries and the country's core competitiveness, and it is the core link connecting the education chain and the industrial chain.

[0003] However, current engineering talent cultivation in universities generally faces the following problems: First, there is a dynamic mismatch between professional structure and industry demand. The adjustment cycle of professional settings and curriculum systems is usually 3-5 years, which lags far behind the 1-2 year iteration speed of emerging industry positions, resulting in a mismatch between graduates' knowledge structure and actual job requirements. Second, homogenized training based on classes suppresses individual differences. The unified curriculum system and evaluation standards cannot take into account the individual differences of students in terms of interests, abilities, and development directions, and lack personalized growth path support for individualized teaching. Third, the matching of people and jobs lacks precision and interpretability. Traditional recruitment and employment guidance rely on resume keywords and human experience, which is difficult to accurately depict the job ability requirements and students' actual ability structure, resulting in the coexistence of difficulty in finding employment and difficulty in recruiting. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings and deficiencies of existing technologies and to provide a method for planning personalized growth paths for college students based on AI big data.

[0005] The technical solution adopted by this invention is as follows: This application provides a method for planning personalized growth paths for college students based on AI big data, including the following steps: constructing a full-link multi-source data system spanning enrollment, training, and employment; through data integration and correlation processing, linking student characteristic data in the enrollment stage, learning process data in the training stage, and job demand data in the employment stage across stages to form a unified data foundation covering the entire student growth cycle; based on the full-link multi-source data system, constructing a capability correlation model, which is used to establish a mapping relationship between student capabilities and job requirements; based on the capability correlation model, generating personalized growth paths for college students, and dynamically adjusting the personalized growth paths according to preset conditions, wherein the personalized growth paths at least include recommendations for development directions that match student capabilities and job requirements.

[0006] In some embodiments, the end-to-end multi-source data system further includes a collection of unstructured documents, which includes training programs, course outlines, job descriptions, industry reports, and policy documents. The data integration and association processing includes: using unified student ID, course ID, and job ID as primary keys to perform cross-system association of enrollment, training, and employment data, and constructing a unified data model oriented towards students, courses, jobs, and competency units.

[0007] In some embodiments, the student characteristics data includes high school academic data, ability assessment data, competition / innovation experience data, career interest data, and personality trait data; The learning process data includes course learning data, practical participation data, and comprehensive quality evaluation data; The job demand data includes job posting information, industry demand trend data, and regional talent shortage data.

[0008] In some embodiments, the capability association model is constructed based on the KAQ capability model, specifically including: A pre-trained language model is used to perform entity recognition and relation extraction on text data in the full-link multi-source data system. Construct a job competency map and a professional competency map. The job competency map defines the mapping relationship and weight between jobs and KAQ competency units, and the professional competency map defines the mapping relationship and weight between graduation requirements, courses, and KAQ competency units.

[0009] In some embodiments, the pre-trained language model is a language model fine-tuned with corpus from the field of engineering education, and the entity recognition and relation extraction are achieved through the pre-trained language model and sequence labeling subnetwork; the job competency graph and professional competency graph are stored in a graph database, and the demand changes of KAQ competency units are monitored on a semester basis. When the change exceeds a preset threshold, the graph is updated.

[0010] In some embodiments, the generation of the personalized growth path includes: Based on the aforementioned ability association model, construct student ability profiles and job ability models; Through a multi-stage decision optimization algorithm, a personalized growth path is generated, including course selection options, practical activity arrangements, and target career cluster recommendations, while meeting the constraints of the training program.

[0011] In some embodiments, the construction of the student competency profile includes the following steps: mapping students' course learning data and practical participation data to scores of corresponding competency units, constructing a time-series competency vector, giving higher weight to the latest learning outcomes by combining a time decay factor, and predicting competency development trends through a time-series model.

[0012] In some embodiments, the construction of the job competency model includes the following steps: encoding the job requirement text, extracting knowledge points, skill requirements and quality requirements and mapping them to competency units, generating a job competency weight vector, calculating the matching degree between the student's competency vector and the job competency weight vector, and recommending career clusters with satisfactory matching degree as growth goals.

[0013] In some embodiments, the preset conditions include: the gap between the student's ability and the target job requirements exceeds a threshold, the student's career interests change significantly, the industry job requirements are significantly adjusted, and the student's learning path is changed; the dynamic adjustment includes: prioritizing the learning resources corresponding to the unmet abilities, replanning the growth path only for the future stage, and retaining the validity of the completed learning content.

[0014] In some embodiments, a knowledge enhancement step is also included: building a RAG layer, performing vectorization processing and retrieval on unstructured documents in the full-link multi-source data system, and using the retrieved relevant document fragments as context to optimize the construction accuracy of the capability association model and the rationality of the generation of personalized growth paths.

[0015] The beneficial effects of this invention are as follows: By constructing a full-link multi-source data system, this invention achieves unified association of data across stages, providing a complete and accurate data source for planning; by establishing a quantitative mapping between students' abilities and job requirements through a capability association model, it solves the problem of low person-job matching; and by adapting to changes in student growth and industry needs through a dynamic adjustment mechanism, it ensures continuous adaptation of the path, ultimately improving the accuracy of talent cultivation in universities and students' employment competitiveness. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention.

[0017] Figure 1 This is a flowchart of the AI-based big data-driven personalized growth path planning method for college students in this invention. Figure 2 This is a framework diagram of the AI-based big data-driven personalized growth path planning method for college students in this invention. Figure 3 This is a collection diagram of student origin characteristic data in this invention; Figure 4 This is a diagram showing the collection of learning process data in this invention; Figure 5This is a collection diagram of job requirement data in this invention. Detailed Implementation

[0018] The following description provides specific application scenarios and requirements for this specification, intended to enable those skilled in the art to make and use the contents of this specification. Various partial modifications to the disclosed embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments and applications without departing from the spirit and scope of this specification. Therefore, this specification is not limited to the embodiments shown, but rather to the widest scope consistent with the claims.

[0019] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "longitudinal", "lateral", "radial", "length", "width", "thickness", "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are mainly for the purpose of better describing this application and its embodiments, and are not intended to limit the indicated device, element or component to have a specific orientation, or to be constructed and operated in a specific orientation.

[0020] It should be noted that the terms "first," "second," and similar words do not indicate any order, quantity, or importance, but are only used to distinguish different components and should not be construed as limiting the embodiments of this application.

[0021] It should be noted that the terms "installation," "setup," "equipped with," "connection," and "connected" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral structures; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium, or internal connections between two devices, components, or parts.

[0022] It should be noted that the terms "in some embodiments," "exemplarily," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "in some embodiments," "exemplarily," or "for example" should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "in some embodiments," "exemplarily," and "for example" is intended to present related concepts in a specific manner, meaning that a particular feature, structure, or characteristic described in connection with an embodiment may include in at least one embodiment of this application. The appearance of the above terms in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0023] Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0024] Regarding the accompanying drawings of this application, it should be clearly understood that the drawings are for illustrative and descriptive purposes only and are not intended to limit the scope of this specification. It should also be understood that the drawings are not necessarily drawn to scale.

[0025] The cultivation of engineering talents in universities is currently facing three core dilemmas: the adjustment of major settings and curriculum systems lags behind the iteration of industry positions, resulting in a disconnect between graduates' knowledge structure and needs; the homogeneous training model based on classes ignores the differences in students' interests and abilities and lacks precise personalized guidance; and the matching of people and jobs relies on resume keywords and human experience, which cannot achieve a precise match between students' abilities and job requirements, resulting in a dual contradiction between employment and recruitment.

[0026] To address the aforementioned problems, this invention proposes a personalized growth path planning method for college students based on AI big data, such as... Figures 1 to 5As shown, the core technical solution of this method revolves around the entire process of "data foundation - capability mapping - path planning," specifically including three core steps: constructing a full-link multi-source data system, constructing a capability association model, and generating and dynamically adjusting personalized growth paths. Specifically, it constructs a full-link multi-source data system spanning enrollment, training, and employment. Through data integration and association processing, it links student characteristic data in the enrollment stage, learning process data in the training stage, and job requirement data in the employment stage across stages, forming a unified data foundation covering the entire student growth cycle. Based on the full-link multi-source data system, a capability association model is constructed to establish the mapping relationship between student capabilities and job requirements. Based on the capability association model, personalized growth paths for university students are generated, and these personalized growth paths are dynamically adjusted according to preset conditions. Each personalized growth path includes at least a development direction recommendation matching the student's capabilities and job requirements. Furthermore, a RAG layer is added to achieve knowledge enhancement.

[0027] In some embodiments, the end-to-end multi-source data system serves as the fundamental support for achieving precise planning throughout the entire lifecycle, breaking down data barriers at each stage of "enrollment-training-employment". When constructing this system, firstly, multi-dimensional data is collected from university enrollment systems, academic affairs management systems, online learning platforms, employment service platforms, and third-party recruitment websites through interface integration and web crawling technology, forming a data source covering the entire student growth cycle. Among these, online learning platforms include MOOCs, Rain Classroom, and XuetangX; employment service platforms include university employment information networks, 51job campus edition, and China University Talent Network; and third-party recruitment websites include Zhaopin.com, BOSS Zhipin, and Liepin Campus.

[0028] The comprehensive multi-source data system includes student characteristic data during the enrollment phase, learning process data during the training phase, and job requirement data during the employment phase. In the data integration and association processing stage, using unified student ID, course ID, and job ID as primary keys, ETL (Extract-Transform-Load) tools are used to clean (remove duplicates and abnormal data), transform (unify data format and measurement units), and load cross-system data, ultimately constructing a unified data model oriented towards "student-course-job-competency unit." For example, the 85 points achieved by student ID 2024001 in the "Data Structures" course (course ID: CS001) is associated with the "Data Structures Knowledge" competency unit for the Algorithm Engineer job (job ID: J0001), providing data support for subsequent competency matching. This data integration method effectively solves the problems of data fragmentation and one-sided planning basis in existing technologies, ensuring the integrity and relevance of the data foundation.

[0029] For example, the student characteristics data includes time series of monthly, mid-term, and final exam scores for each subject in high school, as well as combinations of elective subjects (e.g., physics + chemistry + biology); standardized ability assessment results such as mathematical thinking and logical reasoning (e.g., ACT math scores, scores of engineering ability tests independently set by domestic universities); data on subject competitions and technological innovation experiences (e.g., provincial second prize in the National High School Physics Competition, award records in the Youth Science and Technology Innovation Competition); Holland Career Interest Scale scores (e.g., research-oriented score of 85, realistic score of 60); and main dimension data of the Big Five personality traits (e.g., extraversion of 65, conscientiousness of 90). For unstructured materials such as competition award certificates and personal statements, key information of the certificates is extracted using OCR technology, and NLP technology is used to parse the interest and trait descriptions in the personal statement text and supplement them into structured fields to ensure the comprehensiveness of the data.

[0030] For example, the learning process data includes course learning data (scores for regular assignments, mid-term and final exam scores, and lab operation scores for each course), viewing time on online learning platforms (e.g., 100% completion rate for watching videos in the "Python Programming" course), chapter quiz accuracy (e.g., 92%), practical participation data (scores for lab reports, evaluation of roles and achievements in school-enterprise cooperation projects, records of participation in subject competitions, etc.), and multi-dimensional comprehensive quality evaluation data of teachers on students' collaborative ability and innovative awareness (e.g., 4.5 points for collaborative ability and 4 points for innovative awareness in a 5-point scoring system). This data can reflect the dynamic development of students' abilities in real time and provide a real-time basis for personalized path planning.

[0031] For example, the job demand data includes job information (job responsibilities, qualifications, career development paths) from campus recruitment and third-party platforms, trend reports published by industry associations (such as the "Industrial Software Talent Development Report" published by the China Electronics Technology Standardization Institute), and regional shortage talent directories published by local human resources and social security departments (such as the directory of artificial intelligence operation and maintenance talents in short supply in the Yangtze River Delta region). After being collected through API interfaces and web crawling, the data is cleaned to remove invalid information, forming a standardized job demand dataset, ensuring the accuracy of job demand characterization.

[0032] In some embodiments, the full-link multi-source data system also includes an unstructured document set as a supplement. The core function of the unstructured document set is to provide in-depth knowledge support for the construction of the capability association model, supplement the semantic details of the structured data in capability description and demand interpretation, and at the same time provide policy basis and industry reference for the generation of the personalized growth path, ensuring that the path planning is both in line with educational norms and adapted to industry realities.

[0033] The unstructured document collection includes university engineering program training plans, course outlines, job descriptions (JDs), industry trend reports, and education policy documents. Specifically, the university engineering program training plans are customized programs for each university's specific program; course outlines are teaching guidelines for courses specifically designed for different majors; job descriptions include those for algorithm engineers, embedded development engineers, and data analysts; industry trend reports are professional analysis documents reflecting industry development trends, technological innovation directions, and talent demand characteristics, such as the "White Paper on Talent Demand in the Artificial Intelligence Industry" and the "Report on Talent Development in the Industrial Internet Industry"; and education policy documents are official normative documents guiding educational reform, talent cultivation, and employment and entrepreneurship, such as the "Opinions on Deepening the Reform of the Modern Vocational Education System" and the "Guiding Opinions of the Ministry of Education on Strengthening Employment and Entrepreneurship Work for University Graduates."

[0034] The processing of this unstructured information requires the integration of AI technology to transform it from unstructured to structured, and from semantic information to knowledge elements. A possible processing flow is as follows: First, extract the text content using document parsing tools (such as PDFMiner or Python-Docx), removing invalid information such as headers, footers, and redundant formatting. Then, use a pre-trained language model fine-tuned with corpus data from the engineering education field to perform entity recognition and relation extraction, extracting core relationships such as "course-knowledge unit," "job-skill requirements," and "policy-training objectives." Finally, associate the processed structured knowledge elements with the unified data model of the full-link multi-source data system, while storing the original text fragments in a vector database to provide a data foundation for enhanced retrieval in the RAG layer.

[0035] Based on the aforementioned end-to-end multi-source data system, the capability association model is constructed. This model 2 serves as the core bridge for establishing a quantitative mapping between student capabilities and job requirements, addressing the issue of ambiguous matching between the two in existing technologies. The capability association model is built upon the KAQ (Knowledge-Abilities-Quality) capability model, and its implementation process includes three stages: text processing, graph construction, and graph updating.

[0036] In the text processing stage, a pre-trained language model is used to perform entity recognition and relation extraction on text data from the multi-source data system across the entire chain. This pre-trained language model is a language model fine-tuned from engineering education corpora, such as the QWen-7B model. The fine-tuned corpus includes over 1000 engineering course outlines, over 5000 engineering job descriptions (JDs), and over 200 industry talent demand reports. Accurate processing is achieved through a combination of the pre-trained language model and a sequence labeling subnetwork, which employs a BiLSTM-CRF structure. For example, entities such as "proficient in Python / C++ programming language" and "possess machine learning algorithm optimization capabilities" are extracted from the "algorithm engineer" job description. Simultaneously, relationships such as "job-skill," "course-knowledge," and "knowledge-ability" are identified, providing core elements for graph construction.

[0037] In the graph construction phase, based on the extracted entities and relationships, job competency graphs and professional competency graphs are constructed. The job competency graph defines the mapping relationship and weights between jobs and KAQ competency units. For example, the mapping relationship between the job "Embedded Development Engineer" and "Embedded System Knowledge (weight 0.3)," "C Language Programming Ability (weight 0.4)," and "Rigorous and Meticulous Qualities (weight 0.3)" is given. The professional competency graph defines the mapping relationship and weights between graduation requirements, courses, and KAQ competency units. For example, the mapping relationship between "Automation Major Graduation Requirement 1 (Mastering Automatic Control Principles)" and the "Automatic Control Principles" course (weight 0.6), "Automatic Control Knowledge (weight 0.5)," and "System Modeling Ability (weight 0.5)" is given. The quantitative expression of competency requirements is achieved through weight settings.

[0038] The job competency graph and professional competency graph are stored in a graph database, such as Neo4j, facilitating multi-hop relationship queries and weight calculations. Simultaneously, the competency association model features a dynamic update mechanism. The graph update module monitors the frequency and weight changes of each Key Ability Question (KAQ) unit in the job requirement text on a semester basis. When the change exceeds a preset threshold, the reconstruction of the corresponding graph subgraph is triggered. For example, if the frequency of "large model application ability" in the job description for "Artificial Intelligence Engineer" over the past three months increases by 15% compared to the previous semester (exceeding the 10% threshold), the weight of that competency unit in the job competency graph is updated, and the association weights between the corresponding courses and that competency unit in the professional competency graph are adjusted simultaneously. This ensures that the competency association model always adapts to changes in industry needs, solving the problem of poor model timeliness in existing technologies.

[0039] Based on the aforementioned competency association model, a personalized growth path is generated, and real-time optimization of the path is achieved through a dynamic adjustment unit to ensure its personalization and adaptability. The generation of the personalized growth path includes constructing a student competency profile, constructing a job competency model, and generating the path through a multi-stage decision optimization algorithm.

[0040] The construction of the student competency profile is centered on quantifying competency. The specific steps are as follows: First, students' course learning data is mapped to initial scores for corresponding KAQ competency units through a course-competency mapping table. For example, 88 points in the "Python Programming" course is mapped to 85 points in the "Python Programming Ability" unit. Practical participation data, such as "Provincial First Prize in the Undergraduate Electronic Design Contest," is mapped to 90 points in the "Innovation Ability" unit and 88 points in the "Teamwork Ability" unit. Second, a semester-level time-series competency vector is constructed by combining students' competency scores from each semester. For example, the competency vector for the first semester is: [Data Structure Knowledge 82, Python Programming Ability 85, Innovation Ability 78]. Then, a time decay factor is introduced to give higher weight to the latest learning outcomes, avoiding interference from outdated data in competency assessment. Finally, the time-series competency vector is input into an LSTM time-series model to predict the development trend of key competency units in the next 2-3 semesters. For example, it is predicted that "Machine Learning Ability" will reach 80 points in the third semester. Ultimately, a multi-dimensional student competency profile is formed, including current competency status and future development trends, achieving accurate characterization of student competency.

[0041] The construction of the job competency model is used to achieve quantitative matching of job requirements. The specific steps are as follows: the job requirement text is encoded using the pre-trained language model, knowledge points, skill requirements and quality requirements are extracted and mapped to KAQ competency units, and a job competency weight vector is generated. For example, the job competency weight vector for "Data Analysis Engineer" is: [SQL database knowledge 0.3, data visualization skills 0.4, data analysis thinking 0.3].

[0042] Then, the cosine similarity between the student's ability vector and the job's ability weight vector is calculated as the basic matching degree (e.g., 0.75). A weighted adjustment is then made by combining the matching degree of career interests (e.g., "data science" for the student, matching degree 0.8) and the matching degree of work environment preferences (e.g., a preference for internet companies in first-tier cities for the student, matching degree 0.7). The weights are allocated as follows: basic matching degree 0.5, career interest matching degree 0.3, and work environment preference matching degree 0.2, resulting in a final comprehensive matching degree of 0.74. One to two career clusters with a comprehensive matching degree ≥ 0.7 are recommended as the student's growth goals, ensuring that the target job is highly compatible with the student's abilities and interests.

[0043] The multi-stage decision optimization algorithm is used to generate specific growth path content. The optimal bundle search algorithm, with a bundle width of 4, aims to maximize the comprehensive benefits of job matching, ability achievement, and academic compliance at graduation. Under the premise of meeting the constraints of the training program, such as prerequisites for required courses, a total credit limit of 160 credits, and a minimum practical credit requirement of 15 credits, the algorithm defines "student's ability vector in semester t + completed course set + participated practical project set" as the state, and "course selection in semester t+1 + micro-major selection + competition / practical activity set" as the action. The optimal action sequence is obtained through multi-stage decision calculation. For example, the personalized growth path generated for computer science students is: Semester 1: *Python Programming*, *Advanced Mathematics*; Semester 2: *Data Structures*, *Database Principles* + school-level programming competition; Semester 3: *Machine Learning Fundamentals* + Big Data Analysis Training Project; Semester 4: *Large Model Application Development* + corporate internship (data analyst position); target career cluster: algorithm engineer, achieving a precise connection between courses, practice, and career goals.

[0044] The dynamic adjustment unit is used to monitor path adaptability and trigger adjustments. The preset adjustment conditions include four categories: the student's achievement rate of a certain key KAQ competency unit is lower than 60%-70% (e.g., "Machine Learning Ability" is 72 points in reality, the target is 80 points, the achievement rate is 90% but the gap is 8 points, which exceeds the 5-point threshold); the change in the student's career interest assessment results is ≥30% (e.g., the student changes from "Hardware Development" to "Software Development", and the interest matching degree changes by 35%); the weight of KAQ competency units in industry job requirements changes by ≥15% (e.g., the weight of "CAD Secondary Development Ability" in the "Industrial Software Engineer" position increases from 0.2 to 0.4, a change of 100%); and the student's learning path changes (e.g., changing majors to Artificial Intelligence or joining the "Large Model Application" micro-major). When an adjustment is triggered, the dynamic adjustment unit prioritizes the learning resources corresponding to the unmet abilities. For example, it increases the recommendation weight of the "Deep Learning in Practice" course for students who do not meet the "Machine Learning Ability" standard, re-clusters and recommends target career clusters with declining demand, and only re-plans the path for the next semester, retaining the validity of completed courses / practices to ensure the efficiency and continuity of the path adjustment.

[0045] In some embodiments, the RAG layer is used to optimize model construction accuracy and path generation rationality, reducing model illusion. The RAG layer includes a document processing unit, a vector retrieval unit, and a context injection unit. The specific working process is as follows: the document processing unit performs text segmentation on the unstructured document set and uses a Transformer-based text embedding model such as Sentence-BERT to convert it into a 768-dimensional fixed-dimensional vector; the vector retrieval unit performs semantic similarity retrieval based on vector cosine similarity calculation, and graph structure retrieval based on multi-hop relationship query of "course-ability-job". The retrieval results of the two are fused by weight, such as semantic similarity weight 0.6-0.7 and graph structure retrieval weight 0.3-0.4, to determine the final relevant context fragments; the context injection unit injects the retrieved relevant document fragments (such as the description of the knowledge point "Application of Convolutional Neural Networks" in the "Machine Learning Course Syllabus") into the pre-trained language model, providing accurate knowledge support for the construction of the ability association model and the generation of personalized growth paths, further improving the reliability and accuracy of the method.

[0046] The workflow of this method is as follows: First, data from the entire lifecycle of enrollment, training, and employment is collected and integrated through a multi-source data system to form a unified data foundation. Second, the competency association model, based on the KAQ model, uses text processing and graph construction to achieve a quantitative mapping between student competencies and job requirements. Then, the path planning module constructs student competency profiles and job competency models based on the competency association model, and generates personalized growth paths through a multi-stage decision optimization algorithm. At the same time, the dynamic adjustment unit monitors preset conditions in real time and dynamically optimizes the paths. The RAG layer enhances the accuracy of the model and paths through knowledge augmentation, ultimately achieving a closed-loop process of "data support - competency matching - path generation - dynamic optimization".

[0047] It is understood that ordinary variations and substitutions that can be made by those skilled in the art within the scope of the technical solution of this invention, such as replacing the pre-trained language model with GPT-4 Turbo, or replacing the multi-stage decision optimization algorithm with a genetic algorithm, should also be included within the protection scope of this invention.

[0048] In summary, after reading this detailed disclosure, those skilled in the art will understand that the foregoing detailed disclosure is presented by way of example only and is not restrictive. Although not explicitly stated herein, those skilled in the art will understand that the requirements of this application encompass various reasonable changes, improvements, and modifications to the embodiments. These changes, improvements, and modifications are intended to be made by this application and are within the spirit and scope of the exemplary embodiments of this application.

[0049] Furthermore, it should be understood that in the foregoing description of the embodiments of this application, various features are combined in a single embodiment, drawing, or description for the purpose of simplifying the understanding of a feature. However, this does not mean that the combination of these features is necessary, and those skilled in the art may readily identify some of the devices as separate embodiments when reading this application. That is, the embodiments in this application can also be understood as an integration of multiple sub-embodiments. It is also valid when each sub-embodiment contains fewer than all the features of a single foregoing disclosed embodiment.

[0050] Finally, it should be understood that the embodiments disclosed herein are illustrative of the principles of the embodiments of this application. Other modified embodiments are also within the scope of this application. Therefore, the embodiments disclosed herein are merely examples and not limitations. Those skilled in the art can adopt alternative configurations to implement the applications in this application based on the embodiments in this application. Therefore, the embodiments of this application are not limited to the embodiments precisely described in the application.

Claims

1. A method for planning personalized growth paths for college students based on AI and big data, characterized in that: Includes the following steps: Construct a multi-source data system that spans the entire process of enrollment, training, and employment. Through data integration and correlation processing, link student characteristic data in the enrollment stage, learning process data in the training stage, and job demand data in the employment stage to form a unified data foundation covering the entire student growth cycle. Based on the aforementioned full-link multi-source data system, a capability association model is constructed, which is used to establish the mapping relationship between students' capabilities and job requirements. Based on the aforementioned ability association model, personalized growth paths for college students are generated, and these personalized growth paths are dynamically adjusted according to preset conditions. The personalized growth paths include at least recommendations for development directions that match the students' abilities and job requirements.

2. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, The full-link multi-source data system also includes a collection of unstructured documents, which includes training programs, course outlines, job descriptions, industry reports, and policy documents. The data integration and association processing includes: using unified student ID, course ID, and job ID as primary keys to perform cross-system association of enrollment, training, and employment data, and constructing a unified data model oriented towards students, courses, jobs, and competency units.

3. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, The student characteristics data include high school academic data, ability assessment data, competition / innovation experience data, career interest data, and personality trait data. The learning process data includes course learning data, practical participation data, and comprehensive quality evaluation data; The job demand data includes job posting information, industry demand trend data, and regional talent shortage data.

4. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, The capability association model is built based on the KAQ capability model and specifically includes: A pre-trained language model is used to perform entity recognition and relation extraction on text data in the full-link multi-source data system. Construct a job competency map and a professional competency map. The job competency map defines the mapping relationship and weight between jobs and KAQ competency units, and the professional competency map defines the mapping relationship and weight between graduation requirements, courses, and KAQ competency units.

5. The method for planning personalized growth paths for college students based on AI big data according to claim 4, characterized in that, The pre-trained language model is a language model fine-tuned with corpus from the field of engineering education. The entity recognition and relation extraction are achieved through the pre-trained language model and sequence labeling subnetwork. The job competency graph and professional competency graph are stored in a graph database. The demand changes of KAQ competency units are monitored on a semester basis. When the change exceeds a preset threshold, the graph is updated.

6. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, The generation of the personalized growth path includes: Based on the aforementioned ability association model, construct student ability profiles and job ability models; Through a multi-stage decision optimization algorithm, a personalized growth path is generated, including course selection options, practical activity arrangements, and target career cluster recommendations, while meeting the constraints of the training program.

7. The method for planning personalized growth paths for college students based on AI big data according to claim 6, characterized in that, The construction of the student competency profile includes the following steps: mapping students' course learning data and practical participation data to scores of corresponding competency units, constructing a time-series competency vector, giving higher weight to the latest learning outcomes by combining a time decay factor, and predicting competency development trends through a time-series model.

8. The method for planning personalized growth paths for college students based on AI big data according to claim 6, characterized in that, The construction of the job competency model includes the following steps: encoding the job requirement text, extracting knowledge points, skill requirements and quality requirements and mapping them to competency units, generating a job competency weight vector, calculating the matching degree between the student's competency vector and the job competency weight vector, and recommending career clusters with satisfactory matching degree as growth goals.

9. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, The preset conditions include: the gap between the student's ability and the target job requirements exceeds a threshold, the student's career interests change significantly, the industry job requirements are significantly adjusted, and the student's learning path is changed; the dynamic adjustment includes: prioritizing learning resources corresponding to the unmet abilities, replanning the growth path only for the future stage, and retaining the validity of the completed learning content.

10. The method for planning personalized growth paths for college students based on AI big data according to claim 1, characterized in that, It also includes a knowledge enhancement step: building a RAG layer to perform vectorization processing and retrieval of unstructured documents in the full-link multi-source data system, and using the retrieved relevant document fragments as context to optimize the construction accuracy of the capability association model and the rationality of the generation of personalized growth paths.

Images