Enrollment clue management method
Through the encrypted transmission and the use of the portrait model library, the problem of the inability to accurately evaluate the degree of matching between students and courses in the existing admission clue management is solved, the data security and admission efficiency are improved, and a multi-dimensional course portrait model is built to improve the courses. Intelligent and accurate admissions work.
Patent Information
- Application Number
- CN202410949009.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
Existing methods of enrollment lead management cannot accurately assess student matching with courses and have shortcomings in data security.
By encrypting transmission and storing student information, using the portrait model library and semantic recognition model for data screening and matching, generating matching clue data, and ensuring data security and privacy protection through a unique unlocking key mechanism, while building a multi-dimensional course portrait model To improve matching accuracy.
It achieves a more accurate assessment of student-curricular matching, improves enrollment efficiency and quality, ensures data security and privacy protection, and improves the intelligence and accuracy of enrollment work.
Smart Images

Figure CN118916532A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of clue management, and in particular to an enrollment clue management method. Background Art
[0002] In the expert recruitment process, especially for graduate students, it is crucial to effectively screen and manage leads that meet the recruitment requirements. Traditional recruitment lead management methods mainly rely on manual screening and evaluation, which is inefficient and cannot accurately assess the match between students and specific experts or courses.
[0003] The Chinese patent "A professional enrollment prediction method and system based on a multi-dimensional data model", publication number: CN109472414B, authorization announcement date: September 3, 2021, specifically discloses obtaining each student's growth file, intended college data, student subject selection, grades, historical score lines of various colleges, enrollment numbers and admission rates; filtering by batches to form a data pool, and analyzing the six dimensions of the data pool's students' subject selection, grades, historical high school undergraduate and junior college admission rates, historical college enrollment numbers, admission score lines, and current enrollment policies, and extracting data features from each dimension; cross-extracting the screened student data with the college's historical batch lines and the number of students enrolled in each province to obtain student data that meets the college's requirements, and matching and predicting the corresponding enrollment data for the students according to the enrollment policy. However, this method mainly focuses on hard-condition screening based on grades, and does not take into account flexible conditions such as students' personal interests, specialties, scientific research experience, etc. Therefore, it is insufficient in evaluating the matching degree between students and specific experts or courses. In addition, the aforementioned patent does not clearly mention how to ensure the security of student information during the admissions process. Summary of the invention
[0004] In view of the deficiency that the matching degree between students and courses cannot be accurately evaluated after obtaining the existing enrollment clues, the purpose of the present invention is to provide an enrollment clue management method to more accurately evaluate the matching degree between students and courses, while paying attention to the security protection of student information and improving enrollment efficiency and quality.
[0005] To achieve the above object, the present invention provides the following technical solution: a method for managing enrollment leads, comprising the following steps: S1: Encrypted transmission and acquisition of enrollment clue data, wherein the enrollment clue data includes student information; S2: merging and processing the enrollment clue data according to a preset data screening strategy to generate matching clue data, storing the matching clue data in a closed database, and generating corresponding matching clues; S3: searching for a matching course portrait model in the portrait model library according to the matching clue data, calculating the matching degree between each matching clue and its corresponding course portrait model through a preset clue matching algorithm, and weighting them according to the matching degree to calculate a comprehensive matching result; S4: Determine the potential value level of each clue data based on the comprehensive matching results; S5: Generate a unique unlocking key corresponding to the student information according to different potential value levels, upload the unique unlocking key to the unique unlocking port through a preset information retrieval terminal, and retrieve the corresponding student information in the closed database.
[0006] In this technical solution, the data security during the transmission process is ensured by encrypting the transmission of enrollment clue data, effectively preventing data leakage or illegal interception. At the same time, the student information is stored in a closed database, and the information is retrieved by generating a unique unlocking key, which further enhances the security of the data and the protection of student privacy. The enrollment clue data is merged and processed through the preset data screening strategy to generate more accurate and valuable matching clue data, which not only improves the availability of the data, but also lays the foundation for subsequent accurate matching. Through the clue matching algorithm, the matching degree of each clue with the course portrait model can be accurately calculated, thereby realizing personalized course recommendation. By comprehensively matching the clue data and evaluating the potential value level, the enrollment staff can more intuitively understand the potential value of each clue, so as to give priority to high-value clues and improve the efficiency and success rate of enrollment. At the same time, students are recommended to the most suitable courses for themselves, which also helps to improve students' learning experience and satisfaction. Through the preset information retrieval terminal and the unique unlocking port, the enrollment staff can retrieve the required student information anytime and anywhere without cumbersome procedures or waiting time, which greatly improves the efficiency and quality of enrollment.
[0007] The present invention is further configured as follows: between step S2 and step S3, the following steps are also included to construct a portrait model library: A1: Generate corresponding feature keywords based on the course feature data uploaded by the corresponding course terminal; A2: Associating the characteristic keywords through a preset language association model to generate a word library link list corresponding to each characteristic keyword, each word library link list has a number of associated words corresponding to each characteristic keyword, and each associated word corresponds to a characteristic association value; A3: Associating each associated word with the history student information database to obtain the association information of each associated word, the association information includes the history association feature, the association degree data and the fitness degree data, the association degree data reflects the association degree between the associated word and the history association feature, and the fitness degree data reflects the actual fitness degree between the history association feature and the corresponding course; A4: The course portrait model is generated by associating a word library list according to the associated information. The course portrait model uses the associated words as the index of the associated nodes and the associated information as the content of the associated nodes.
[0008] In this technical solution, in step A1, according to the rich data uploaded by the course terminal, the course features are accurately extracted, and then the corresponding feature keywords are generated, which ensures that the construction of the portrait model starts from the core features of the course and lays a solid foundation for the subsequent steps. Subsequently, in step A2, these feature keywords are deeply associated using the preset language association model, which not only generates a vocabulary list corresponding to each feature keyword, but also assigns a feature association value to each associated word, thereby enriching the dimension and depth of the portrait model, so that it can more comprehensively reflect the characteristics of the course. In step A3, each associated word is accurately matched with the historical student information database to obtain detailed association information of each associated word, which includes historical association features, association data, and fitness data, providing us with in-depth insights into the fitness of courses and students. The association data and fitness data accurately reflect the closeness between the associated words and the historical association features, as well as the actual matching degree of these features with the corresponding courses. Finally, in step A4, the course portrait model is carefully constructed using this rich association information. This model uses related words as indexes and related information as content to form a three-dimensional, multi-dimensional course description, which not only provides a comprehensive course perspective, but also brings unprecedented convenience and accuracy to enrollment work. By implementing this enrollment lead management method, a series of refined processes are introduced between steps S2 and S3, aiming to build a comprehensive and in-depth portrait model library, thereby improving the intelligence and accuracy of enrollment work.
[0009] The present invention is further configured as follows: the data screening strategy extracts the data content of each enrollment clue data to input into a preset semantic recognition model, and modifies the data content through the semantic recognition model to generate enrollment semantic data, formats and normalizes the enrollment semantic data to obtain a number of enrollment semantic items and corresponding enrollment semantic fields, and calculates the confidence value of the corresponding enrollment semantic field through a preset confidence calculation algorithm for the same enrollment semantic item, so as to screen the enrollment semantic field with a confidence value higher than the preset one to generate the matching clue data.
[0010] In this technical solution, the enrollment clue management method significantly improves the efficiency and accuracy of enrollment work through a refined data screening strategy. First, the data content of each enrollment clue is extracted and input into a preset semantic recognition model. Through this model, the data content can be corrected and improved to generate more accurate enrollment semantic data. Next, the system will format and normalize these enrollment semantic data, decomposing them into multiple enrollment semantic items and corresponding enrollment semantic fields, making the data more structured and convenient for subsequent analysis and processing. By calculating the same enrollment semantic items, the system can generate a confidence value for each enrollment semantic field, reflecting the accuracy and credibility of the field, and providing strong data support for enrollment personnel. Finally, the system will screen out enrollment semantic fields with confidence values higher than the preset standard according to the preset threshold, and generate the so-called matching clue data. These data are not only highly accurate, but also highly targeted, which can greatly improve the efficiency and quality of enrollment work.
[0011] The present invention is further configured as follows: the confidence calculation algorithm is Among them, P is the confidence value of one of the enrollment semantic fields corresponding to the enrollment semantic item, x i is the information association value of the i-th enrollment semantic field with the same enrollment semantic item, u i is the semantic correction deviation of the i-th enrollment semantic field with the same enrollment semantic item, I is the total number of enrollment semantic fields with the same enrollment semantic item, y j is the information association value of the jth enrollment semantic field with different enrollment semantic items, q j is the semantic correction deviation of the jth enrollment semantic field with different enrollment semantic items, J is the total number of enrollment semantic fields with different enrollment semantic items, is the preset weight value of the same category, is the preset heterogeneous weight value.
[0012] In this technical solution, the confidence calculation algorithm calculates an accurate confidence value for each enrollment semantic field by comprehensively considering multiple factors, including the information association value and semantic correction deviation of enrollment semantic fields with the same and different enrollment semantic items. In this algorithm, P represents the confidence value of a certain enrollment semantic field corresponding to the enrollment semantic item. This value is obtained through a complex calculation formula, which takes into account the information x of the enrollment semantic fields with the same enrollment semantic item. i 、u i and information y of the enrollment semantic field with different enrollment semantic items j ,q j In this way, the algorithm can comprehensively evaluate the accuracy and credibility of each field. In addition, the algorithm also introduces preset weight values of the same category. and heterogeneous weights These two parameters allow the algorithm to weight different types of admissions semantic fields differently when calculating confidence values, thereby more accurately reflecting the importance of each field in the admissions leads.
[0013] The present invention is further configured as follows: the clue matching algorithm matches the matching clues with the keywords in the course portrait model to determine the activated matching path in the course portrait model, and weights the matching path through a preset matching association function combined with the association data and the fitness data to calculate the matching degree corresponding to the matching path.
[0014] In this technical solution, the clue matching algorithm can quickly determine the activated matching path in the course portrait model by accurately matching the matching clues with the keywords in the course portrait model. It is not only highly intelligent, but also greatly improves the accuracy and efficiency of matching. After determining the activated matching path, the preset matching association function is further used to weight these matching paths in combination with the association data and the fitness data. In this way, the algorithm can comprehensively consider multiple factors and calculate an accurate matching degree for each matching path, which helps admissions staff to have a more comprehensive understanding of the matching degree between clues and courses.
[0015] The present invention is further configured as follows: each course corresponds to a preset level index table, the level index table stores a number of comprehensive result ranges and corresponding potential value levels, and the corresponding potential value level is retrieved according to the comprehensive result range into which the comprehensive matching result falls.
[0016] In this technical solution, the enrollment clue management method presets a detailed level index table for each course. This index table carefully divides multiple comprehensive result ranges and sets corresponding potential value levels for each range. When the enrollment staff obtains the comprehensive matching results, the system can quickly retrieve the corresponding potential value level from the level index table according to the range of this result. It not only greatly simplifies the value assessment process of enrollment clues, but also makes the assessment results more objective and accurate. Enrollment staff no longer need to rely on subjective judgment or complex calculations to determine the potential value of the clues. They only need to check the range in which the comprehensive matching results fall to quickly understand the potential value level corresponding to the clue. The efficiency and accuracy of enrollment work have been significantly improved. Enrollment staff can focus more on clues with high potential value, optimize resource allocation, and improve enrollment conversion rate. At the same time, this also provides strong support for the formulation and adjustment of enrollment strategies, which helps to achieve the refinement and scientificization of enrollment work.
[0017] The present invention is further configured as follows: the unique unlocking key is configured with a passive triggering class, an active triggering class, an authorized use class, and an authorized index class corresponding to different potential value levels; The passive trigger type unique unlocking key is configured to be generated by encrypting a closed random factor generated inside a closed database. When the unique unlocking key retrieves data in the corresponding closed database, the corresponding closed verification factor in the closed database changes to inactivate the unique unlocking key. The unique unlocking key of the active trigger type is configured to be generated by encrypting a control random factor generated by the system background, and the system background can inactivate the unique unlocking key by modifying the corresponding control verification factor; The unique unlocking key of the authorized use class is configured to be generated by encrypting the authorization random factor generated by the authorization terminal, and the authorization terminal can enable the unique unlocking key to be activated for a preset first authorization time by editing the corresponding authorization verification factor; The unique unlocking key of the authorized index class is configured to be generated by encrypting the index random factor generated by the authorized terminal. When the unique unlocking key retrieves the data of the corresponding closed database, the index random factor is replaced by the solid address factor to generate a new unique decryption key. The unique decryption key generated by the solid address factor can obtain the data of the corresponding address of the closed database.
[0018] In this technical solution, the enrollment clue management method provides refined and multi-level access control for enrollment clues of different potential values through different types of unique unlocking keys. The security of the passive trigger type unique unlocking key lies in that it will automatically deactivate after each use. This key is encrypted and generated by a closed random factor generated inside a closed database. Once the data is retrieved, the corresponding closed verification factor will change, making the key invalid, effectively preventing data leakage and illegal access. The active trigger type unique unlocking key provides more flexibility. This type of key is controlled by the system background. The key can be manually deactivated by modifying the control verification factor, which is convenient for managers to manage keys according to actual needs, such as quickly closing the access rights of a key when necessary. The authorized use type unique unlocking key is for a specific authorized terminal. This type of key can be encrypted and generated by the authorized random factor generated by the authorized terminal, and the key can be activated by editing the authorized verification factor and setting the first authorization time, which not only ensures the security of the data, but also facilitates specific users to access the data within the specified time. The unique unlocking key of the authorized index class provides a more complex access mechanism. After use, this type of key will generate a new decryption key. The new decryption key is generated by a solid address factor and can be used to obtain data at a specific address in a closed database, which not only ensures the security of the data, but also provides a flexible and traceable data access method.
[0019] The present invention is further configured as follows: after step S5, the following steps are further included: Using the clue mining model, we can deeply mine and analyze the integrated information of students, discover their strengths and the matching degree with the course, and label the clues according to the potential value level; Dynamically adjust tags and matching levels based on actual development of leads; If the clue data shows that the match between the student and the course has improved, the potential value level of the label will be increased; If the lead data shows that the student's match with the course is decreasing, the potential value level of the lead label can be lowered accordingly or the label can be changed to reflect the new student needs.
[0020] In this technical solution, by using the clue mining model to deeply mine and analyze the integrated information of students, it is possible to more accurately discover the students' strengths and their matching degree with specific courses, providing valuable reference information for admissions personnel. In addition, this method is also highly dynamic and flexible. According to the actual development of the clues, such as student feedback, communication information with students, and student behavior changes, the clue labels and matching evaluations are adjusted in real time and dynamically, ensuring the timeliness and accuracy of admissions information, so that admissions personnel can respond quickly to market changes and changes in student needs. When the clue data shows that the matching degree between students and courses has improved, the method will correspondingly increase the potential value level of its label, thereby helping admissions personnel give priority to those students who are highly matched with the courses. On the contrary, if the clue data shows that the matching degree between students and courses has decreased, the method will promptly reduce the potential value level of the clue label or change the label to accurately reflect the new needs and intentions of students.
[0021] The present invention is further configured as follows: the clue mining model includes: Integrate and cleanse student information collected from various channels; Key features are extracted from student information, including age, gender, region, academic performance, and interests and hobbies; based on the extracted features, students are accurately classified and predicted by using machine learning algorithms, and potential student clues are identified through model training, and divided into different potential value levels, and the intrinsic connections and patterns between different features are discovered.
[0022] In this technical solution, firstly, through the data integration and cleaning link, the student information collected from various channels can be effectively integrated and deeply cleaned to ensure that the processed data is accurate, consistent and reliable, laying a solid foundation for subsequent data analysis and avoiding analysis bias caused by data errors or inconsistencies. Secondly, the feature extraction link enables the model to accurately extract key features from student information, such as age, gender, region, academic performance, interests and hobbies, etc. These features not only fully reflect the basic situation and personality characteristics of students, but also provide strong data support for subsequent prediction and classification. In the prediction and classification link, the model uses advanced machine learning algorithms to accurately classify and predict students based on the extracted features. After model training, potential student clues can be accurately identified and divided into different levels according to their potential value, which improves the pertinence and efficiency of enrollment work and helps educational institutions quickly lock in high-quality enrollment targets. Finally, the association analysis link enables the model to deeply explore the internal connections and laws between different features, which not only helps to understand the characteristics and needs of student groups more comprehensively, but also provides strong data support for educational institutions to formulate more accurate enrollment strategies.
[0023] The present invention is further configured as follows: the encrypted transmission is presented through a client, and the client includes a teacher terminal and a student terminal; on the student terminal, the student's own student information and clue tags related to the student are viewed; on the teacher terminal, all clue tags are viewed through identity authentication.
[0024] In this technical solution, students can view their own student information through the student end, which helps students confirm and verify the accuracy of personal information, and promptly discover and correct possible errors. Students can view clue tags related to themselves. Clue tags are generated based on students' learning interests, intended courses, interactive behaviors and other data, so that students can more clearly understand their position in the enrollment system and possible learning recommendations. Through the teacher end, teachers can view all clue data after passing identity authentication, providing teachers with comprehensive enrollment information and student profiles, which helps teachers better understand the needs and intentions of potential students, so as to formulate more accurate enrollment strategies. Through centralized clue data management, teachers can track and manage potential students more efficiently, reduce the time for information search and sorting, and improve work efficiency. The clue tags displayed on the teacher end can help teachers quickly identify students' characteristics and needs, so as to conduct more targeted enrollment communication and follow-up.
[0025] The advantages of the present invention are: (1) it can more accurately evaluate the matching degree between students and courses, while paying attention to the security protection of student information, thereby improving the efficiency and quality of enrollment; (2) it can ensure the security of data during transmission by encrypting the transmission of enrollment clue data, effectively preventing data leakage or illegal interception. At the same time, it can store student information in a closed database and retrieve the information by generating a unique unlocking key, thereby further enhancing the security of data and the protection of student privacy. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 The present invention is a flow chart of an enrollment lead management method. DETAILED DESCRIPTION
[0027] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0028] The present invention will be further described below in conjunction with the accompanying drawings.
[0029] Embodiment 1: This embodiment provides a method for managing enrollment leads, including the following steps: S1: Encrypted transmission and acquisition of enrollment clue data, wherein the enrollment clue data includes student information; S2: merging and processing the enrollment clue data according to a preset data screening strategy to generate matching clue data, storing the matching clue data in a closed database, and generating corresponding matching clues; S3: searching for a matching course portrait model in the portrait model library according to the matching clue data, calculating the matching degree between each matching clue and its corresponding course portrait model through a preset clue matching algorithm, and weighting them according to the matching degree to calculate a comprehensive matching result; S4: Determine the potential value level of each clue data based on the comprehensive matching results; S5: Generate a unique unlocking key corresponding to the student information according to different potential value levels, upload the unique unlocking key to the unique unlocking port through a preset information retrieval terminal, and retrieve the corresponding student information in the closed database.
[0030] In this embodiment, by encrypting and transmitting the enrollment clue data, the security of the data during the transmission process is ensured, and data leakage or illegal interception is effectively prevented. At the same time, the student information is stored in a closed database, and the information is retrieved by generating a unique unlocking key, which further enhances the security of the data and the protection of student privacy. The enrollment clue data is merged and processed by a preset data screening strategy to generate more accurate and valuable matching clue data, which not only improves the availability of the data, but also lays the foundation for subsequent accurate matching. Through the clue matching algorithm, the matching degree of each clue with the course portrait model can be accurately calculated, thereby realizing personalized course recommendation. By comprehensively matching and rating the potential value of the clue data, the enrollment staff can more intuitively understand the potential value of each clue, thereby giving priority to high-value clues and improving the efficiency and success rate of enrollment. At the same time, students are recommended to the most suitable courses for themselves, which also helps to improve students' learning experience and satisfaction. Through the preset information retrieval terminal and the unique unlocking port, the enrollment staff can retrieve the required student information anytime and anywhere without cumbersome procedures or waiting time, which greatly improves the efficiency and quality of enrollment.
[0031] refer to Figure 1 , between step S2 and step S3, the following steps are also included to build a portrait model library: A1: Generate corresponding feature keywords based on the course feature data uploaded by the corresponding course terminal; A2: Associating the characteristic keywords through a preset language association model to generate a word library link list corresponding to each characteristic keyword, each word library link list has a number of associated words corresponding to each characteristic keyword, and each associated word corresponds to a characteristic association value; A3: Associating each associated word with the history student information database to obtain the association information of each associated word, the association information includes the history association feature, the association degree data and the fitness degree data, the association degree data reflects the association degree between the associated word and the history association feature, and the fitness degree data reflects the actual fitness degree between the history association feature and the corresponding course; A4: The course portrait model is generated by associating a word library list according to the associated information. The course portrait model uses the associated words as the index of the associated nodes and the associated information as the content of the associated nodes.
[0032] In this technical solution, in step A1, according to the rich data uploaded by the course terminal, the course features are accurately extracted, and then the corresponding feature keywords are generated, which ensures that the construction of the portrait model starts from the core features of the course and lays a solid foundation for the subsequent steps. Subsequently, in step A2, these feature keywords are deeply associated using the preset language association model, which not only generates a vocabulary list corresponding to each feature keyword, but also assigns a feature association value to each associated word, thereby enriching the dimension and depth of the portrait model, so that it can more comprehensively reflect the characteristics of the course. In step A3, each associated word is accurately matched with the historical student information database to obtain detailed association information of each associated word, which includes historical association features, association data, and fitness data, providing us with in-depth insights into the fitness of courses and students. The association data and fitness data accurately reflect the closeness between the associated words and the historical association features, as well as the actual matching degree of these features with the corresponding courses. Finally, in step A4, the course portrait model is carefully constructed using this rich association information. This model uses related words as indexes and related information as content to form a three-dimensional, multi-dimensional course description, which not only provides a comprehensive course perspective, but also brings unprecedented convenience and accuracy to enrollment work. By implementing this enrollment lead management method, a series of refined processes are introduced between steps S2 and S3, aiming to build a comprehensive and in-depth portrait model library, thereby improving the intelligence and accuracy of enrollment work.
[0033] refer to Figure 1 The data screening strategy extracts the data content of each enrollment clue data to input into a preset semantic recognition model, and modifies the data content through the semantic recognition model to generate enrollment semantic data, formats and normalizes the enrollment semantic data to obtain a number of enrollment semantic items and corresponding enrollment semantic fields, and calculates the confidence value of the corresponding enrollment semantic field for the same enrollment semantic item through a preset confidence calculation algorithm, so as to screen the enrollment semantic field with a confidence value higher than the preset one to generate the matching clue data.
[0034] In this technical solution, the enrollment clue management method significantly improves the efficiency and accuracy of enrollment work through a refined data screening strategy. First, the data content of each enrollment clue is extracted and input into a preset semantic recognition model. Through this model, the data content can be corrected and improved to generate more accurate enrollment semantic data. Next, the system will format and normalize these enrollment semantic data, decomposing them into multiple enrollment semantic items and corresponding enrollment semantic fields, making the data more structured and convenient for subsequent analysis and processing. By calculating the same enrollment semantic items, the system can generate a confidence value for each enrollment semantic field, reflecting the accuracy and credibility of the field, and providing strong data support for enrollment personnel. Finally, the system will screen out enrollment semantic fields with confidence values higher than the preset standard according to the preset threshold, and generate the so-called matching clue data. These data are not only highly accurate, but also highly targeted, which can greatly improve the efficiency and quality of enrollment work.
[0035] Preferably, the confidence calculation algorithm is Among them, P is the confidence value of one of the enrollment semantic fields corresponding to the enrollment semantic item, x i is the information association value of the i-th enrollment semantic field with the same enrollment semantic item, u i is the semantic correction deviation of the i-th enrollment semantic field with the same enrollment semantic item, I is the total number of enrollment semantic fields with the same enrollment semantic item, y j is the information association value of the jth enrollment semantic field with different enrollment semantic items, q j is the semantic correction deviation of the jth enrollment semantic field with different enrollment semantic items, J is the total number of enrollment semantic fields with different enrollment semantic items, is the preset weight value of the same category, is the preset heterogeneous weight value.
[0036] In this technical solution, the confidence calculation algorithm calculates an accurate confidence value for each enrollment semantic field by comprehensively considering multiple factors, including the information association value and semantic correction deviation of enrollment semantic fields with the same and different enrollment semantic items. In this algorithm, P represents the confidence value of a certain enrollment semantic field corresponding to the enrollment semantic item. This value is obtained through a complex calculation formula, which takes into account the information x of the enrollment semantic fields with the same enrollment semantic item. i 、u i and information y of the enrollment semantic field with different enrollment semantic items j ,q j In this way, the algorithm can comprehensively evaluate the accuracy and credibility of each field. In addition, the algorithm also introduces preset weight values of the same category. and heterogeneous weights These two parameters allow the algorithm to weight different types of admissions semantic fields differently when calculating confidence values, thereby more accurately reflecting the importance of each field in the admissions leads.
[0037] In one embodiment of the present invention, the clue matching algorithm matches the matching clues with keywords in the course portrait model to determine the activated matching path in the course portrait model, and weights the matching path by combining the association data and the fitness data through a preset matching association function to calculate the matching degree corresponding to the matching path.
[0038] In this technical solution, the clue matching algorithm can quickly determine the activated matching path in the course portrait model by accurately matching the matching clues with the keywords in the course portrait model. It is not only highly intelligent, but also greatly improves the accuracy and efficiency of matching. After determining the activated matching path, the preset matching association function is further used to weight these matching paths in combination with the association data and the fitness data. In this way, the algorithm can comprehensively consider multiple factors and calculate an accurate matching degree for each matching path, which helps admissions staff to have a more comprehensive understanding of the matching degree between clues and courses.
[0039] In one embodiment of the present invention, reference Figure 1 Each course has a preset level index table corresponding to it. The level index table stores several comprehensive result ranges and corresponding potential value levels. The corresponding potential value level is retrieved according to the comprehensive result range into which the comprehensive matching result falls.
[0040] In this technical solution, the enrollment clue management method presets a detailed level index table for each course. This index table carefully divides multiple comprehensive result ranges and sets corresponding potential value levels for each range. When the enrollment staff obtains the comprehensive matching results, the system can quickly retrieve the corresponding potential value level from the level index table according to the range of this result. It not only greatly simplifies the value assessment process of enrollment clues, but also makes the assessment results more objective and accurate. Enrollment staff no longer need to rely on subjective judgment or complex calculations to determine the potential value of the clues. They only need to check the range in which the comprehensive matching results fall to quickly understand the potential value level corresponding to the clue. The efficiency and accuracy of enrollment work have been significantly improved. Enrollment staff can focus more on clues with high potential value, optimize resource allocation, and improve enrollment conversion rate. At the same time, this also provides strong support for the formulation and adjustment of enrollment strategies, which helps to achieve the refinement and scientificization of enrollment work.
[0041] In one embodiment of the present invention, reference Figure 1, the unique unlocking key is configured with passive triggering class, active triggering class, authorized use class, and authorized index class corresponding to different potential value levels; The passive trigger type unique unlocking key is configured to be generated by encrypting a closed random factor generated inside a closed database. When the unique unlocking key retrieves data in the corresponding closed database, the corresponding closed verification factor in the closed database changes to inactivate the unique unlocking key. The unique unlocking key of the active trigger type is configured to be generated by encrypting a control random factor generated by the system background, and the system background can inactivate the unique unlocking key by modifying the corresponding control verification factor; The unique unlocking key of the authorized use class is configured to be generated by encrypting the authorization random factor generated by the authorization terminal, and the authorization terminal can enable the unique unlocking key to be activated for a preset first authorization time by editing the corresponding authorization verification factor; The unique unlocking key of the authorized index class is configured to be generated by encrypting the index random factor generated by the authorized terminal. When the unique unlocking key retrieves the data of the corresponding closed database, the index random factor is replaced by the solid address factor to generate a new unique decryption key. The unique decryption key generated by the solid address factor can obtain the data of the corresponding address of the closed database.
[0042] In this technical solution, the enrollment clue management method provides refined and multi-level access control for enrollment clues of different potential values through different types of unique unlocking keys. The security of the passive trigger type unique unlocking key lies in that it will automatically deactivate after each use. This key is encrypted and generated by a closed random factor generated inside a closed database. Once the data is retrieved, the corresponding closed verification factor will change, making the key invalid, effectively preventing data leakage and illegal access. The active trigger type unique unlocking key provides more flexibility. This type of key is controlled by the system background. The key can be manually deactivated by modifying the control verification factor, which is convenient for managers to manage keys according to actual needs, such as quickly closing the access rights of a key when necessary. The authorized use type unique unlocking key is for a specific authorized terminal. This type of key can be encrypted and generated by the authorized random factor generated by the authorized terminal, and the key can be activated by editing the authorized verification factor and setting the first authorization time, which not only ensures the security of the data, but also facilitates specific users to access the data within the specified time. The unique unlocking key of the authorized index class provides a more complex access mechanism. After use, this type of key will generate a new decryption key. The new decryption key is generated by a solid address factor and can be used to obtain data at a specific address in a closed database, which not only ensures the security of the data, but also provides a flexible and traceable data access method.
[0043] In one embodiment of the present invention, reference Figure 1 , after step S5, further comprising: Using the clue mining model, we can deeply mine and analyze the integrated information of students, discover their strengths and the matching degree with the course, and label the clues according to the potential value level; Dynamically adjust tags and matching levels based on actual development of leads; If the clue data shows that the match between the student and the course has improved, the potential value level of the label will be increased; If the clue data shows that the match between the student and the course is reduced, the potential value level of the clue label is reduced or the label is changed to reflect the new student needs. By using the clue mining model to deeply mine and analyze the integrated information of students, it is possible to more accurately discover the student's strengths and their match with specific courses, providing valuable reference information for admissions staff. In addition, the method is highly dynamic and flexible. According to the actual development of the clue, such as student feedback, communication information with students, and student behavior changes, the clue label and match evaluation are adjusted in real time and dynamically, ensuring the timeliness and accuracy of admissions information, allowing admissions staff to respond quickly to market changes and changes in student needs. When the clue data shows that the match between the student and the course is improved, the method will correspondingly increase the potential value level of its label, thereby helping admissions staff to give priority to those students who are highly matched with the course. On the contrary, if the clue data shows that the match between the student and the course is reduced, the method will promptly reduce the potential value level of the clue label or change the label to accurately reflect the student's new needs and intentions.
[0044] In one embodiment of the present invention, reference Figure 1 , the clue mining model includes: Integrate and cleanse student information collected from various channels; Extract key features from student information, including age, gender, region, academic performance, and hobbies; based on the extracted features, use machine learning algorithms to accurately classify and predict students, identify potential student clues through model training, and classify them into different potential value levels, and discover the internal connections and rules between different features. First, through the data integration and cleaning link, it is possible to effectively integrate and deeply clean the student information collected from various channels to ensure that the processed data is accurate, consistent and reliable, laying a solid foundation for subsequent data analysis and avoiding analysis bias caused by data errors or inconsistencies. Secondly, the feature extraction link enables the model to accurately extract key features from student information, such as age, gender, region, academic performance, hobbies, etc. These features not only fully reflect the basic situation and personality characteristics of students, but also provide strong data support for subsequent prediction and classification. In the prediction and classification link, the model uses advanced machine learning algorithms to accurately classify and predict students based on the extracted features. After model training, potential student leads can be accurately identified and divided into different levels according to their potential value, which improves the pertinence and efficiency of enrollment work and helps educational institutions quickly lock in high-quality enrollment targets. Finally, the association analysis link enables the model to deeply explore the internal connections and rules between different features, which not only helps to more comprehensively understand the characteristics and needs of student groups, but also provides strong data support for educational institutions to formulate more accurate enrollment strategies.
[0045] Preferably, the encrypted transmission is presented through a client, and the client includes a teacher terminal and a student terminal; on the student terminal, the student information and the clue tags related to the student are viewed; on the teacher terminal, all clue tags are viewed through identity verification. Students can view their own student information through the student terminal, which helps students confirm and verify the accuracy of personal information, timely discover and correct possible errors, and students can view clue tags related to themselves. Clue tags are generated based on students' learning interests, intended courses, interactive behaviors and other data, so that students can more clearly understand their positioning in the enrollment system and possible learning recommendations. Through the teacher terminal, teachers can view all clue data after identity verification, providing teachers with comprehensive enrollment information and student profiles, helping teachers to better understand the needs and intentions of potential students, and thus formulate more accurate enrollment strategies. Through centralized clue data management, teachers can track and manage potential students more efficiently, reduce the time for information search and sorting, and improve work efficiency. The clue tags displayed on the teacher terminal can help teachers quickly identify the characteristics and needs of students, so as to conduct more targeted enrollment communication and follow-up.
[0046] Understandably, student information includes name, gender, mobile phone number, ID number, birthday, personal email address, intended campus, application stage, application grade, admissions source, admissions counselor, graduation year, current school, current grade, etc.
[0047] It is understandable that the ways to obtain enrollment lead data include but are not limited to official website placement, briefing sessions, etc. to attract parents to fill out enrollment forms.
[0048] Understandably, when assessing the duplication of clue data, including checking for duplicate student ID numbers, checking for duplicate student mobile numbers, and checking for duplicate student mobile numbers + student names, when clues with system duplication or high similarity are found, they are merged. At the same time, when continuing to enter clues with the same rules at a later stage, the system will prompt to reduce the entry of invalid clues.
[0049] Preferably, the certificate type and certificate number fields are set as required.
[0050] Preferably, the encrypted transmission includes: Acquire multiple plaintext clue data groups of clue data to be encrypted, and encrypt the multiple plaintext clue data groups respectively based on the key and the student ID number to obtain multiple ciphertext clue data groups; The clue data to be encrypted includes at least the student ID number and mobile phone number; Determine the anti-replay factor of the current data transmission process, wherein the anti-replay factors of two adjacent data transmission processes are different; generate an authentication code according to the multiple ciphertext clue data groups and the anti-replay factor, including: using the multiple ciphertext clue data groups as inputs of the CBC mode, and using the anti-replay factor as the initial vector of the CBC mode; based on the key combined with the student number, obtain multiple pseudo ciphertext clue data groups through iterative encryption operations in the CBC mode, and use the last pseudo ciphertext clue data group as the authentication code; Each of the multiple ciphertext clue data groups is split and reassembled according to preset rules to generate multiple reassembled ciphertext clue data groups, wherein the preset rules are predetermined by the data sending end and the data receiving end; the multiple reassembled ciphertext clue data groups, the anti-replay factor and the authentication code are sent to the data receiving end, wherein the authentication code is used to compare with the verification code generated by the data receiving end to identify whether the data has been tampered with, and the verification code is generated by the data receiving end based on the multiple ciphertext clue data groups and the anti-replay factor determined by the multiple reassembled ciphertext clue data groups.
[0051] In this technical solution, by encrypting students' sensitive information, including student ID numbers, mobile phone numbers, etc., the security of data during transmission is ensured. The encryption process uses an encryption algorithm based on a key and student ID number, making it difficult to decrypt the data even if it is intercepted, thereby effectively protecting students' privacy. By generating an authentication code and verifying it at the data receiving end, it is possible to identify whether the data has been tampered with during transmission, thereby ensuring the integrity and authenticity of the data and avoiding errors or confusion in enrollment information caused by data tampering. By grouping and encrypting the plaintext clue data, and then splitting and reorganizing the ciphertext clue data, the method not only improves the efficiency of data transmission, but also increases the accuracy of data verification at the data receiving end. At the same time, the iterative encryption operation using the CBC mode further enhances the security of the data. By introducing an anti-replay factor, it is ensured that the anti-replay factors of two adjacent data transmission processes are different, thereby effectively preventing replay attacks, making each data transmission unique, and even if the attacker intercepts the previous data packet, it cannot be reused in subsequent transmissions. Both the data sender and the data receiver follow the same preset rules to split, reorganize and verify the data, ensuring that both parties can maintain consistency when processing data and avoiding information confusion or loss caused by different data processing methods. Through mechanisms such as encrypted transmission, data verification and anti-replay attack protection, the security and reliability of the enrollment clue management method are significantly enhanced, which not only protects students' privacy information, but also ensures the accuracy and integrity of enrollment data, providing a strong guarantee for the smooth progress of enrollment work.
[0052] Preferably, the intention information includes at least: source of leads, communication information, registration intention, and behavioral performance information; the sources of leads include recommendations from old students with high intention, proactive consultations with medium intention, and advertising campaigns with low intention. This setting helps the admissions team to develop more targeted follow-up strategies based on leads from different sources. By distinguishing the intention of leads, the admissions team can allocate resources and time more efficiently. Understanding the source and intention of leads helps the admissions team develop more precise marketing strategies. Attaching importance to recommendations from old students not only shows respect and gratitude to old students, but also helps to improve their satisfaction and loyalty. Word-of-mouth communication often brings in higher-quality potential students.
[0053] Preferably, high-intent referral leads from old students can be processed first and followed up to improve conversion rate. For low-intent leads from advertising campaigns, the quality and intent of the leads can be improved by optimizing the advertising content or positioning.
[0054] Preferably, the communication information includes communication depth and communication frequency, and the communication strategy is adjusted according to the communication information. By considering the communication depth and communication frequency, the admissions staff can more accurately understand the needs and concerns of potential students, thereby adjusting the communication content to make it closer to the actual situation of the students, thereby improving the communication effect. According to the feedback and interaction of the students, the communication strategy can be flexibly adjusted, so that the students can feel more attention and personalized services, thereby enhancing the students' stickiness and trust in the admissions agency or school. By analyzing the communication information, the admissions team can more reasonably allocate human and time resources. Among them, for potential students with deeper communication depth and higher frequency, more energy and resources can be invested in follow-up to improve the conversion rate. According to the feedback of the communication information, timely adjustment of the strategy can help eliminate the concerns of potential students and enhance their willingness to sign up, thereby improving the admissions conversion rate. Continuously collecting and analyzing communication information and adjusting the communication strategy based on this information will help the admissions team to continuously improve and optimize the entire admissions process, making it more efficient and accurate.
[0055] Preferably, the registration intention includes clear registration intention and potential registration intention, and the students with clear registration intention are followed up, and the students with potential registration intention are further communicated. By timely following up on students with clear registration intention, it is ensured that these high-intention students receive sufficient attention and guidance, so as to facilitate their registration decision more easily. It also helps the admissions team to quickly lock in and convert potential students who have shown strong interest. For students with potential registration intention, further communication can deepen their understanding of courses or services, stimulate their interest, and then convert them into students with clear registration intention. It helps the admissions team to continuously expand its enrollment base and attract more potential students. By distinguishing between students with clear registration intention and potential registration intention, the admissions team can allocate resources and time more reasonably.
[0056] It is understandable that the behavioral performance information includes the frequency of participation in institution-related activities and the frequency of active request for information. By collecting information on the frequency of students' participation in institution activities and the frequency of request for information, the admissions team can more accurately understand the students' interests, needs, and activity. This helps to build a more accurate student portrait and provide data support for subsequent enrollment strategies and personalized services. Understanding the frequency of students' participation in activities can help the admissions team determine which types of activities are more popular with students, so as to adjust and optimize future marketing activities. At the same time, based on the frequency of students' active request for information, it can be determined what content they are more interested in, and then provide more targeted information and materials.
[0057] The specific description of the present invention in the above embodiments is only used to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Technical engineers in this field may make some non-essential improvements and adjustments to the present invention based on the contents of the above invention, which fall within the scope of protection of the present invention.
Claims
1. A method for managing enrollment leads, characterized in that: The following steps are involved: S1: Encrypted transmission and acquisition of enrollment clue data, wherein the enrollment clue data includes student information; S2: merging and processing the enrollment clue data according to a preset data screening strategy to generate matching clue data, storing the matching clue data in a closed database, and generating corresponding matching clues; S3: searching for a matching course portrait model in the portrait model library according to the matching clue data, calculating the matching degree between each matching clue and its corresponding course portrait model through a preset clue matching algorithm, and weighting them according to the matching degree to calculate a comprehensive matching result; S4: Determine the potential value level of each clue data based on the comprehensive matching results; S5: Generate a unique unlocking key corresponding to the student information according to different potential value levels, upload the unique unlocking key to the unique unlocking port through a preset information retrieval terminal, and retrieve the corresponding student information in the closed database.
2. A recruitment lead management method according to claim 1, characterized in that: Between step S2 and step S3, the following steps are also included to build a portrait model library: A1: Generate corresponding feature keywords based on the course feature data uploaded by the corresponding course terminal; A2: The characteristic keywords are associated with each other through a preset language association model to generate a word library link list corresponding to each characteristic keyword. Each word library link list has a number of associated words corresponding to each characteristic keyword, and each associated word corresponds to a characteristic association value; A3: Associating each associated word with the history student information database to obtain associated information of each associated word, the associated information includes historical associated features, associated degree data and fit data, the associated degree data reflects the degree of association between the associated word and the historical associated features, and the fit data reflects the actual fit between the historical associated features and the corresponding courses; A4: The course portrait model is generated by associating a word library list according to the associated information. The course portrait model uses the associated words as the index of the associated nodes and the associated information as the content of the associated nodes.
3. The method for managing enrollment leads according to claim 1, characterized in that: The data screening strategy extracts the data content of each enrollment clue data to input into a preset semantic recognition model, and modifies the data content through the semantic recognition model to generate enrollment semantic data, formats and normalizes the enrollment semantic data to obtain a number of enrollment semantic items and corresponding enrollment semantic fields, and calculates the confidence value of the corresponding enrollment semantic field for the same enrollment semantic item through a preset confidence calculation algorithm, so as to screen the enrollment semantic field with a confidence value higher than the preset one to generate the matching clue data.
4. A recruitment clue management method according to claim 3, characterized in that: The confidence calculation algorithm is Among them, P is the confidence value of one of the enrollment semantic fields corresponding to the enrollment semantic item, x i is the information association value of the i-th enrollment semantic field with the same enrollment semantic item, u i is the semantic correction deviation of the i-th enrollment semantic field with the same enrollment semantic item, I is the total number of enrollment semantic fields with the same enrollment semantic item, y j is the information association value of the jth enrollment semantic field with different enrollment semantic items, q j is the semantic correction deviation of the jth enrollment semantic field with different enrollment semantic items, J is the total number of enrollment semantic fields with different enrollment semantic items, is the preset weight value of the same category, is the preset heterogeneous weight value.
5. The method for managing enrollment leads according to claim 1, characterized in that: The clue matching algorithm matches the matching clues with the keywords in the course portrait model to determine the activated matching path in the course portrait model. The matching path is weighted by combining the association data and the fitness data through a preset matching association function to calculate the matching degree corresponding to the matching path.
6. The method for managing enrollment leads according to claim 4, characterized in that: Each course corresponds to a preset level index table, which stores a number of comprehensive result ranges and corresponding potential value levels. The corresponding potential value level is retrieved according to the comprehensive result range into which the comprehensive matching result falls.
7. A recruitment clue management method according to claim 6, characterized in that: The unique unlocking key is configured with passive triggering class, active triggering class, authorized use class, and authorized index class corresponding to different potential value levels; The passive trigger type unique unlocking key is configured to be generated by encrypting a closed random factor generated inside a closed database. When the unique unlocking key retrieves data in the corresponding closed database, the corresponding closed verification factor in the closed database changes to inactivate the unique unlocking key. The unique unlocking key of the active trigger type is configured to be generated by encrypting a control random factor generated by the system background, and the system background can inactivate the unique unlocking key by modifying the corresponding control verification factor; The unique unlocking key of the authorized use class is configured to be generated by encrypting the authorization random factor generated by the authorization terminal, and the authorization terminal can enable the unique unlocking key to be activated for a preset first authorization time by editing the corresponding authorization verification factor; The unique unlocking key of the authorized index class is configured to be generated by encrypting the index random factor generated by the authorized terminal. When the unique unlocking key retrieves the data of the corresponding closed database, the index random factor is replaced by the solid address factor to generate a new unique decryption key. The unique decryption key generated by the solid address factor can obtain the data of the corresponding address of the closed database.
8. The method for managing enrollment leads according to claim 7, characterized in that: After step S5, the method further includes: using a clue mining model to deeply mine and analyze the integrated information of students, discovering the students' strengths and the degree of matching with the course, and labeling the clues according to the potential value level; Dynamically adjust tags and matching levels based on actual development of leads; If the clue data shows that the match between the student and the course has improved, the potential value level of the label will be increased; If the lead data shows that the student's match with the course is decreasing, the potential value level of the lead label can be lowered accordingly or the label can be changed to reflect the new student needs.
9. A recruitment lead management method according to claim 8, characterized in that: The clue mining model includes: Integrate and cleanse student information collected from various channels; Extract key features from the student information, including age, gender, region, academic performance, and hobbies; Based on the extracted features, students are accurately classified and predicted by using machine learning algorithms. Through model training, potential student clues are identified and divided into different potential value levels, and the intrinsic connections and rules between different features are discovered.
10. A recruitment lead management method according to claim 9, characterized in that: After the encrypted transmission, the information is presented through the client, which includes a teacher terminal and a student terminal. On the student terminal, the student can view his / her own student information and clue tags related to himself / herself. On the teacher terminal, all clue tags can be viewed through identity authentication.
Citation Information
Cited By
Sales clue score design method suitable for user value analysis and medium
CN119762005A
A sales lead scoring design method and medium suitable for user value analysis
CN119762005B
Intelligent enrollment and increase management method and system based on cross verification
CN121981690A
