Online language learning resource intelligent matching method based on artificial intelligence

Abstract

The invention discloses an online language learning resource intelligent matching method based on artificial intelligence, and relates to the technical field of data management. The method comprises the steps of constructing an online database, obtaining resource features and evaluation coefficients of various resource data in the online database, obtaining learning coefficients of users, generating initial matching resources in combination with the evaluation coefficients, obtaining tendency coefficients of the users, constructing a personal database according to the tendency coefficients, and obtaining resource feature ratios of the users. Optimized matching resources are generated in combination with a personal database, corresponding resource matching models are constructed according to different learning coefficients and tendency coefficients and the optimized matching resources thereof, and the optimized matching resources are updated by utilizing the resource matching models; through the technical scheme of the invention, a targeted matching mechanism is favorably formed for the user, intelligent matching of resource data is realized, and the efficiency of online language learning resource matching is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to an online language learning resource intelligent matching method based on artificial intelligence. Background Art

[0002] Intelligent matching of online language learning resources mainly relies on artificial intelligence technologies, including natural language processing, database construction, and deep learning. These methods can comprehensively and deeply analyze learners' needs and characteristics, thereby providing them with targeted learning resources to help learners master language skills faster and better, thereby improving the efficiency of language learning;

[0003] In the prior art, most of the matching mechanisms for resource data lack pertinence and fail to provide users with appropriate resource data based on their learning abilities and learning tendencies. In addition, in the prior art, providing users with resource matching based on all resource data is a technical means with a large amount of computation, which inevitably leads to low efficiency of resource matching. The prior art lacks technical solutions to these problems. In view of the shortcomings of the prior art, the present invention provides an intelligent matching method for online language learning resources based on artificial intelligence. Summary of the invention

[0004] The purpose of the present invention is to provide an online language learning resource intelligent matching method based on artificial intelligence.

[0005] The purpose of the present invention can be achieved by the following technical solution: an online language learning resource intelligent matching method based on artificial intelligence, comprising the following steps:

[0006] Step S1: construct an online database, obtain resource characteristics of each resource data in the online database, and obtain the evaluation coefficient corresponding to each resource data;

[0007] Step S2: Obtain the user's historical learning data and its corresponding learning coefficient, and generate initial matching resources for the user in combination with the evaluation coefficient;

[0008] Step S3: obtaining the user's real-time learning data of the initial matching resources, and obtaining the tendency coefficients of different resource features in the real-time learning data;

[0009] Step S4: constructing a personal database of the user according to the propensity coefficient, establishing a data transmission channel between the personal database and the online database, obtaining the resource feature ratio of the user, and generating optimized matching resources for the user in combination with the personal database;

[0010] Step S5: According to different learning coefficients and inclination coefficients and their optimized matching resources, a corresponding resource matching model is constructed, and the resource matching model is used to update the optimized matching resources.

[0011] Furthermore, the process of building an online database includes:

[0012] Collecting resource data, wherein the resource data refers to various data related to language learning, including basic learning resources, multimedia resources, interactive learning resources, tools and applications;

[0013] Digital processing of various resource data, where digital processing refers to the process of converting non-digital data and processing methods into digital form and processing them using digital technology;

[0014] A corresponding online database is constructed according to the subordinate relationship between the digitally processed resource data, wherein the subordinate relationship refers to the inherent corresponding relationship when the resource data is collected.

[0015] Furthermore, the process of obtaining resource characteristics of each resource data in the online database and obtaining the evaluation coefficient corresponding to each resource data includes:

[0016] The resource features refer to the feature information that distinguishes different resource data from each other, including name features, language features, form features, interaction features, and auxiliary features;

[0017] Set corresponding weight parameters for different resource characteristics, and obtain corresponding evaluation coefficients Z based on the data packet size of each resource data;

[0018]

[0019] Among them, B q is the data packet size of resource data after digital processing, B h The difficulty level of resource data preset, α i is the weight parameter of the resource feature contained in the resource data, i=1, 2, ..., n, n is the number of resource features contained in the resource data, ∈ and γ are both fixed values ​​greater than zero.

[0020] Furthermore, the process of obtaining the user's historical learning data and its corresponding learning coefficient and generating initial matching resources for the user in combination with the evaluation coefficient includes:

[0021] Collect users' personal information and generate personal accounts based on the personal information. Users use the personal accounts to access various resource data on the platform.

[0022] The historical learning data refers to the resource data that the user has used in the platform. The evaluation coefficient corresponding to each historical learning data is recorded as Z j , where j = 1, 2, ..., m, m is the number of historical learning data, and the learning coefficient R of the user is obtained;

[0023]

[0024] The evaluation coefficient of each resource data in the online database is obtained, and each resource data whose evaluation coefficient is less than or equal to the learning coefficient is used as the candidate resource data for the user. A refresh cycle is set. Every time a refresh cycle is reached, a number of resource data are randomly obtained from the candidate resource data as initial matching resources, and they are fed back to the user.

[0025] Furthermore, the process of obtaining the user's real-time learning data of the initial matching resources and obtaining the inclination coefficients of different resource features in the real-time learning data includes:

[0026] Setting an analysis period to obtain real-time learning data of users within a single analysis period, wherein the real-time learning data refers to various initial matching resources used by the user within the analysis period;

[0027] Obtain various real-time learning data containing the same resource characteristics within the analysis cycle, and calculate the evaluation coefficient Z corresponding to each real-time learning data. k , obtain the tendency coefficient Q of the resource characteristics;

[0028]

[0029] Wherein, k=1, 2, ..., b, b is the number of real-time learning data containing the resource feature in the analysis period, μ and τ are both fixed values ​​greater than zero, and the tendency coefficients of other resource features in the analysis period are obtained respectively.

[0030] Furthermore, a personal database of the user is constructed according to the tendency coefficient, and a process of establishing a data transmission channel between the personal database and the online database includes:

[0031] Build a personal database, obtain the tendency coefficients of various resource characteristics in the most recent analysis period, and use various resource characteristics and their tendency coefficients as the entry criteria for the personal database in the current analysis period;

[0032] The entry criteria are used to determine whether resource data in the online database can be entered into the personal database. If all resource features of the resource data intersect with the resource features included in the entry criteria, the resource data is determined to meet the entry criteria.

[0033] A data transmission channel is established between the personal database and the online database, and resource data in the online database that meets the entry criteria is incorporated into the user's personal database through the data transmission channel.

[0034] Furthermore, the process of obtaining the resource feature ratio of the user and generating optimized matching resources for the user in combination with the personal database includes:

[0035] The ratio between the inclination coefficients of various resource characteristics in the user's most recent analysis period is used as the resource characteristic ratio in the current analysis period;

[0036] Each time an analysis cycle is reached, the user's learning coefficient and inclination coefficient are obtained, and the personal database is continuously updated according to the resource characteristics corresponding to each inclination coefficient. The candidate resource data is obtained in the personal database of the current analysis cycle according to the learning coefficient;

[0037] At each refresh cycle, a number of candidate resource data are randomly obtained from the personal database of the current analysis cycle as optimized matching resources, and fed back to the user. The ratio between the number of occurrences of each resource feature in the optimized matching resources generated in each refresh cycle is the same as the resource feature ratio.

[0038] Furthermore, according to different learning coefficients and inclination coefficients and their optimized matching resources, a corresponding resource matching model is constructed, and the process of using the resource matching model to update the optimized matching resources includes:

[0039] Generate a resource matching set according to the learning coefficients and tendency coefficients of different refresh cycles and their corresponding optimized matching resources, and divide the resource matching set into a training set and a test set;

[0040] Construct a convolutional neural network, use different learning coefficients and tendency coefficients in the training set as input data of the convolutional neural network, use the corresponding optimized matching resources in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network;

[0041] The initial convolutional neural network is model verified using the test set, and the initial convolutional neural network with a test error threshold less than or equal to the preset value is output as the resource matching model.

[0042] The learning coefficient and the tendency coefficient of each refresh cycle are continuously obtained and input into the resource matching model to obtain the corresponding optimized matching resources, so as to realize the continuous updating of the optimized matching resources.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. The present invention can provide users with initial matching resources of corresponding difficulty and size according to their learning ability by extracting different resource features from resource data and obtaining the evaluation coefficient of resource data and the learning coefficient of users. The invention can obtain the tendency coefficients of different resource features according to the real-time learning data of users, which can reflect the resource data that users tend to use. Based on this, the user's personal database is constructed. According to the ratio between the tendency coefficients and the personal database, corresponding optimized matching resources are generated for users, which is conducive to forming a targeted matching mechanism for users.

[0045] 2. The present invention continuously obtains the user's learning coefficient and tendency coefficient, and can continuously update the user's personal database according to the resource characteristics corresponding to each tendency coefficient. The optimized matching resources are continuously updated according to the learning coefficient combined with the personal database of the current analysis cycle, and a resource matching model is constructed on this basis, which is conducive to realizing intelligent matching of resource data and can effectively improve the efficiency of online language learning resource matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0047] like Figure 1 As shown, the online language learning resource intelligent matching method based on artificial intelligence includes the following steps:

[0048] Step S1: construct an online database, obtain resource characteristics of each resource data in the online database, and obtain the evaluation coefficient corresponding to each resource data;

[0049] Step S2: Obtain the user's historical learning data and its corresponding learning coefficient, and generate initial matching resources for the user in combination with the evaluation coefficient;

[0050] Step S3: obtaining the user's real-time learning data of the initial matching resources, and obtaining the tendency coefficients of different resource features in the real-time learning data;

[0051] Step S4: constructing a personal database of the user according to the propensity coefficient, establishing a data transmission channel between the personal database and the online database, obtaining the resource feature ratio of the user, and generating optimized matching resources for the user in combination with the personal database;

[0052] Step S5: According to different learning coefficients and inclination coefficients and their optimized matching resources, a corresponding resource matching model is constructed, and the resource matching model is used to update the optimized matching resources.

[0053] It should be further explained that, in the specific implementation process, the process of building an online database includes:

[0054] Collecting resource data, wherein the resource data refers to various data related to language learning, including basic learning resources, multimedia resources, interactive learning resources, tools and applications;

[0055] The basic learning resources include vocabulary lists, grammar guides, pronunciation guides, listening materials, and reading materials; the multimedia resources include video tutorials, audio files, electronic books and periodicals, and reading materials; the interactive learning resources include live courses, language exchanges, forums and communities, practice questions, and test papers; the tools and applications include dictionaries, translation software, speech recognition software, and learning progress tracking software;

[0056] Digital processing is performed on the collected resource data, and the digital processing refers to the process of converting non-digital data and processing methods into digital form and processing them using digital technology;

[0057] For example, converting text, images, audio, and video into digital form and storing the converted data in digital devices to replace paper archives and physical media storage;

[0058] The corresponding online database is constructed according to the subordinate relationship between the digitally processed resource data. The subordinate relationship refers to the corresponding relationship inherent in the collection of resource data. For example, vocabulary lists and listening materials are subordinate to basic learning resources, online courses and exercises are subordinate to interactive learning resources, and audio files and translation software have no subordinate relationship.

[0059] It should be further explained that, in the specific implementation process, the process of obtaining the resource characteristics of each resource data in the online database and obtaining the evaluation coefficient corresponding to each resource data includes:

[0060] The resource features refer to the feature information that distinguishes different resource data from each other, including name features, language features, form features, interaction features, and auxiliary features;

[0061] The name feature is used to distinguish the name differences of different resource data, such as vocabulary, audio files, live courses, translation software, the language feature is used to distinguish the language differences between different resource data, such as English, French, Chinese, Russian, and the form feature is used to distinguish the form differences of different resource data during collection, such as text, image, audio, video;

[0062] The interactive feature is used to distinguish whether different resource data requires the user to interact with others or tools, for example, dictionaries and language communication require the user to interact, while vocabulary lists and audio files do not. The auxiliary feature is used to distinguish whether different resource data can provide auxiliary functions for users, for example, translation software and speech recognition software can provide auxiliary functions for users, while exercises, electronic books and reading materials require users to learn by themselves.

[0063] Set corresponding weight parameters for different resource characteristics, and obtain corresponding evaluation coefficients based on the data packet size of each resource data, denoted as Z;

[0064]

[0065] Among them, B q is the data packet size of resource data after digital processing, B h The difficulty level preset for the resource data, which is used to reflect the language level test level corresponding to the resource data. i is the weight parameter of the resource feature contained in the resource data, i=1, 2, ..., n, n is the number of resource features contained in the resource data, ∈ and γ are both fixed values ​​greater than zero.

[0066] It should be further explained that, in the specific implementation process, the process of obtaining the user's historical learning data and its corresponding learning coefficient, and generating initial matching resources for the user in combination with the evaluation coefficient includes:

[0067] Collect the user's personal information, including name, gender, age, occupation, mobile phone number, ID number, and ID photo, and generate a corresponding personal account for the user based on the personal information. The user needs to use the personal account to use various resource data on the platform;

[0068] The historical learning data refers to the resource data that the user has used in the platform, and the evaluation coefficient Z corresponding to each historical learning data is obtained. j , where j = 1, 2, ..., m, where m is the number of historical learning data, based on which the learning coefficient of the user is obtained, denoted as R;

[0069]

[0070] The evaluation coefficients of various resource data in the online database are obtained, and the resource data whose evaluation coefficients are less than or equal to the learning coefficients are used as candidate resource data. A refresh cycle is set. Every time a refresh cycle is reached, a number of resource data are randomly obtained from the candidate resource data as initial matching resources, and they are fed back to the user.

[0071] It should be further explained that, in the specific implementation process, the process of obtaining the user's real-time learning data of the initial matching resources and obtaining the inclination coefficients of different resource features in the real-time learning data includes:

[0072] Setting an analysis period, where the analysis period is equal to a number of refresh periods, and obtaining real-time learning data of the user for all initial matching resources within a single analysis period, where the real-time learning data refers to various initial matching resources used by the user within the analysis period;

[0073] Obtain various real-time learning data containing the same resource characteristics within the analysis cycle, and calculate the evaluation coefficient Z corresponding to each real-time learning data. k , obtain the tendency coefficient of the resource characteristics, denoted as Q;

[0074]

[0075] Wherein, k=1, 2, ..., b, b is the number of real-time learning data containing the resource feature in the analysis period, μ and τ are both fixed values ​​greater than zero, and the same method is adopted to obtain the tendency coefficients of other resource features in the analysis period.

[0076] It should be further explained that, in the specific implementation process, the process of building a user's personal database based on the propensity coefficient and establishing a data transmission channel between the personal database and the online database includes:

[0077] Build a personal database for the user, obtain the tendency coefficients of various resource characteristics in the most recent analysis cycle, and use the obtained resource characteristics and their corresponding tendency coefficients as the entry criteria for the personal database in the current analysis cycle;

[0078] The entry criteria are used to determine whether resource data in the online database can be entered into the user's personal database, and a corresponding data transmission channel is established between the personal database and the online database, and resource data in the online database that meets the entry criteria is entered into the user's personal database through the data transmission channel;

[0079] Taking any resource data as an example, if all resource characteristics of the resource data have an intersection with the resource characteristics included in the storage criteria, then the resource data is judged to meet the storage criteria; if there is no intersection, then it is judged not to meet.

[0080] It should be further explained that, in the specific implementation process, the process of obtaining the resource feature ratio of the user and generating optimized matching resources for the user in combination with the personal database includes:

[0081] The ratio between the inclination coefficients of various resource features in the user's most recent analysis period is used as the resource feature ratio in the current analysis period, and the corresponding optimized matching resources are generated for the user based on the resource feature ratio and the personal database;

[0082] Each time an analysis cycle is reached, the user's learning coefficient and inclination coefficient are obtained, and the personal database is continuously updated according to the resource characteristics corresponding to each inclination coefficient. The candidate resource data is obtained in the personal database of the current analysis cycle according to the learning coefficient;

[0083] At each refresh cycle, a number of candidate resource data are randomly obtained from the personal database of the current analysis cycle as optimized matching resources, and fed back to the user. The ratio between the number of occurrences of each resource feature in the optimized matching resources generated in each refresh cycle is the same as the resource feature ratio.

[0084] It should be further explained that, in the specific implementation process, according to different learning coefficients and inclination coefficients and their optimized matching resources, a corresponding resource matching model is constructed, and the process of using the resource matching model to update the optimized matching resources includes:

[0085] Generate a resource matching set according to the learning coefficients and the tendency coefficients of different refresh cycles and their corresponding optimized matching resources, and divide the obtained resource matching set into a training set and a test set;

[0086] Construct a convolutional neural network, use different learning coefficients and tendency coefficients in the training set as input data of the convolutional neural network, use the corresponding optimized matching resources in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network;

[0087] The initial convolutional neural network is model verified using the test set, and the initial convolutional neural network with a test error threshold value less than or equal to the preset value is output as the corresponding resource matching model.

[0088] The learning coefficient and the tendency coefficient of each refresh cycle are continuously obtained and input into the resource matching model, and the corresponding optimized matching resources are output through the resource matching model to realize the continuous updating of the optimized matching resources.

[0089] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An online language learning resource intelligent matching method based on artificial intelligence, characterized in that: The following steps are involved: Step S1: construct an online database, obtain resource characteristics of each resource data in the online database, and obtain the evaluation coefficient corresponding to each resource data; Step S2: Obtain the user's historical learning data and its corresponding learning coefficient, and generate initial matching resources for the user in combination with the evaluation coefficient; Step S3: obtaining the user's real-time learning data of the initial matching resources, and obtaining the tendency coefficients of different resource features in the real-time learning data; Step S4: constructing a personal database of the user according to the propensity coefficient, establishing a data transmission channel between the personal database and the online database, obtaining the resource feature ratio of the user, and generating optimized matching resources for the user in combination with the personal database; Step S5: According to different learning coefficients and inclination coefficients and their optimized matching resources, a corresponding resource matching model is constructed, and the resource matching model is used to update the optimized matching resources.

2. The method for intelligently matching online language learning resources based on artificial intelligence according to claim 1, characterized in that: The process of building an online database includes: Collecting resource data, wherein the resource data refers to various data related to language learning, including basic learning resources, multimedia resources, interactive learning resources, tools and applications; Digital processing of various resource data, where digital processing refers to the process of converting non-digital data and processing methods into digital form and processing them using digital technology; A corresponding online database is constructed according to the subordinate relationship between the digitally processed resource data, wherein the subordinate relationship refers to the inherent corresponding relationship when the resource data is collected.

3. The method for intelligent matching of online language learning resources based on artificial intelligence according to claim 2, characterized in that: The process of obtaining resource characteristics and evaluation coefficients for each resource data includes: The resource features refer to the feature information that distinguishes different resource data from each other, including name features, language features, form features, interaction features, and auxiliary features; Set corresponding weight parameters for different resource characteristics, and obtain corresponding evaluation coefficients Z based on the data packet size of each resource data; Among them, B q is the data packet size of resource data after digital processing, B h The difficulty level of resource data preset, α i is the weight parameter of the resource feature contained in the resource data, i=1, 2, ..., n, n is the number of resource features contained in the resource data, ∈ and γ are both fixed values ​​greater than zero.

4. The method for intelligently matching online language learning resources based on artificial intelligence according to claim 3 is characterized in that: The process of obtaining historical learning data and its learning coefficients and generating initial matching resources includes: Collect users' personal information and generate personal accounts based on the personal information. Users use the personal accounts to access various resource data on the platform. The historical learning data refers to the resource data that the user has used in the platform. The evaluation coefficient corresponding to each historical learning data is recorded as Z j , where j = 1, 2, ..., m, m is the number of historical learning data, and the learning coefficient R of the user is obtained; The evaluation coefficient of each resource data in the online database is obtained, and each resource data whose evaluation coefficient is less than or equal to the learning coefficient is used as the candidate resource data for the user. A refresh cycle is set. Every time a refresh cycle is reached, a number of resource data are randomly obtained from the candidate resource data as initial matching resources, and they are fed back to the user.

5. The method for intelligently matching online language learning resources based on artificial intelligence according to claim 4 is characterized in that: The process of obtaining the real-time learning data and its propensity coefficients for different resource characteristics includes: Setting an analysis period to obtain real-time learning data of users within a single analysis period, wherein the real-time learning data refers to various initial matching resources used by the user within the analysis period; Obtain various real-time learning data containing the same resource characteristics within the analysis cycle, and calculate the evaluation coefficient Z corresponding to each real-time learning data. k , obtain the tendency coefficient Q of the resource characteristics; Wherein, k=1, 2, ..., b, b is the number of real-time learning data containing the resource feature in the analysis period, μ and τ are both fixed values ​​greater than zero, and the tendency coefficients of other resource features in the analysis period are obtained respectively.

6. The method for intelligent matching of online language learning resources based on artificial intelligence according to claim 5, characterized in that: The process of building a user's personal database based on the propensity coefficient includes: Build a personal database, obtain the tendency coefficients of various resource characteristics in the most recent analysis period, and use various resource characteristics and their tendency coefficients as the entry criteria for the personal database in the current analysis period; The entry criteria are used to determine whether resource data in the online database can be entered into the personal database. If all resource features of the resource data intersect with the resource features included in the entry criteria, the resource data is determined to meet the entry criteria. A data transmission channel is established between the personal database and the online database, and resource data in the online database that meets the entry criteria is incorporated into the user's personal database through the data transmission channel.

7. The method for intelligently matching online language learning resources based on artificial intelligence according to claim 6, characterized in that: The process of obtaining resource feature ratios and generating optimized matching resources in combination with a personal database includes: The ratio between the inclination coefficients of various resource characteristics in the user's most recent analysis period is used as the resource characteristic ratio in the current analysis period; Each time an analysis cycle is reached, the user's learning coefficient and inclination coefficient are obtained, and the personal database is continuously updated according to the resource characteristics corresponding to each inclination coefficient. The candidate resource data is obtained in the personal database of the current analysis cycle according to the learning coefficient; At each refresh cycle, a number of candidate resource data are randomly obtained from the personal database of the current analysis cycle as optimized matching resources, and fed back to the user. The ratio between the number of occurrences of each resource feature in the optimized matching resources generated in each refresh cycle is the same as the resource feature ratio.

8. The method for intelligent matching of online language learning resources based on artificial intelligence according to claim 7, characterized in that: The process of building a resource matching model and updating and optimizing matching resources includes: Generate a resource matching set according to the learning coefficients and tendency coefficients of different refresh cycles and their corresponding optimized matching resources, and divide the resource matching set into a training set and a test set; Construct a convolutional neural network, use different learning coefficients and tendency coefficients in the training set as input data of the convolutional neural network, use the corresponding optimized matching resources in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The initial convolutional neural network is model verified using the test set, and the initial convolutional neural network with a test error threshold less than or equal to the preset value is output as the resource matching model. The learning coefficient and the tendency coefficient of each refresh cycle are continuously obtained and input into the resource matching model to obtain the corresponding optimized matching resources, so as to realize the continuous updating of the optimized matching resources.

Images