Personalized course intelligent recommendation management platform based on big data

Abstract

The invention relates to the technical field of intelligent curriculum recommendation, and discloses a personalized intelligent curriculum recommendation management platform based on big data, and the platform comprises a user data collection end which is used for obtaining the input information, historical teaching information and teaching test information of a user; the user demand analysis end is used for carrying out user demand analysis according to the input information of the user, the historical teaching information and the teaching test information, and obtaining user demand tags and demand values corresponding to different demand tags; the teaching content label extraction unit is used for extracting the teaching content, obtaining a label of the extracted content, and obtaining a corresponding recommendation value according to the historical playing data of the extracted content and the user evaluation data; and the matching recommendation module is used for matching the user demand tag and the demand values corresponding to different demand tags with the tag of the teaching content and the corresponding recommendation values to obtain the optimal recommendation content of the user. Adaptive recommendation can be performed according to the demand selection of the user, and the recommendation accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent course recommendation, and in particular to a personalized course intelligent recommendation management platform based on big data. Background Art

[0002] In the era of information explosion, online education platforms have brought together a vast amount of course resources, providing users with unprecedented learning opportunities. However, faced with a wide range of courses, users often fall into "choice difficulties" and find it difficult to quickly find the learning content that best suits their needs and interests. With the rapid development of big data technology, it is possible to efficiently collect, store, process and analyze user learning behavior data, providing technical support for building accurate user portraits, and then mining user learning rules from massive data, achieving accurate matching between users and courses, and improving the accuracy of recommendation results and user satisfaction.

[0003] The existing personalized course recommendation system mainly implements the recommendation process through user tag matching. It extracts the corresponding demand tags based on the user's demand data, and then matches them according to the tags of the teaching videos on the course platform, selects the content with a high degree of tag matching, and then recommends it based on the data of the teaching video itself, thereby realizing the matching process according to user needs.

[0004] However, in the process of recommendation, the existing personalized course recommendation system is affected by the subjectivity of the user when making demand judgments directly based on the information input by the user, so the demand information cannot be completely aligned with the user's actual needs. At the same time, although the label comparison method can realize the matching process, the proportion of different demand labels will vary greatly due to different users. Therefore, although the existing personalized course recommendation system can realize the recommendation process, there is room for improvement in the accuracy of the recommendation and the adaptability to different users. Summary of the invention

[0005] In order to improve the accuracy of recommendations and the adaptability to different users, this application provides a personalized course intelligent recommendation management platform based on big data.

[0006] In the first aspect, the present application provides a personalized course intelligent recommendation management platform based on big data, which adopts the following technical solutions:

[0007] A personalized course intelligent recommendation management platform based on big data, including:

[0008] User data collection terminal, used to obtain user input information, historical teaching information and teaching test information;

[0009] The user demand analysis terminal is used to analyze user demand based on user input information, historical teaching information and teaching test information, and obtain user demand tags and demand values ​​corresponding to different demand tags;

[0010] A teaching content label extraction unit is used to extract teaching content, obtain labels of the extracted content, and obtain corresponding recommendation values ​​according to historical playback data and user evaluation data of the extracted content;

[0011] The matching recommendation module is used to match the user's demand labels and the demand values ​​corresponding to different demand labels with the labels of the teaching content and the corresponding recommendation values ​​to obtain the best recommended content for the user.

[0012] By adopting the above technical scheme, compared with the existing personalized course intelligent recommendation system, this embodiment analyzes user needs based on the user's input information, historical teaching information and teaching test information, obtains user demand labels, and also determines the demand values ​​corresponding to different demand labels. The demand value is not only determined based on the user's input information to meet the user's subjective needs, but also objectively analyzed based on the historical teaching information and teaching test information to meet the user's objective needs. Therefore, the obtained demand value can accurately reflect the user's real objective demand status for the teaching course. In addition, this embodiment extracts the teaching content, obtains the extracted content label, and also obtains the corresponding recommendation value based on the historical playback data and user evaluation data of the extracted content. Compared with the direct label matching method, the label and the corresponding recommendation value matching method can better select more accurate and suitable content for recommendation according to the user's needs, thereby improving the accuracy of the recommendation and the adaptability of different users.

[0013] Optionally, the process of analyzing user needs includes:

[0014] Label the user's input information based on big data to obtain the initial demand labels;

[0015] Obtain the user's wrong question information and its corresponding label in the teaching test information, count the corresponding labels of the wrong question information, compare the corresponding labels after counting with the user's historical teaching information, and obtain the comparison result;

[0016] Determine the user demand label and its corresponding demand value based on the initial demand label, the corresponding label after statistics and the comparison result.

[0017] By adopting the above technical scheme, the user's wrong question information and its corresponding labels in the teaching test information are obtained, the corresponding labels of the wrong question information are counted, and the corresponding labels after counting are compared with the user's historical teaching information to obtain the comparison results, wherein the corresponding labels of the wrong question information can be obtained by pre-marking, and can be directly obtained according to the corresponding wrong question information. The wrong question information can also be analyzed based on big data, and its corresponding labels can be obtained according to the analysis results. There is no limitation here. Afterwards, the user's demand label and its corresponding demand value are determined according to the initial demand label, the corresponding label after counting and the comparison result, and then the user's real objective demand status for the teaching course can be accurately reflected according to the obtained demand value, thereby improving the accuracy of recommendation and the adaptability of different users.

[0018] Optionally, the process of determining the user requirement label and its corresponding requirement value includes:

[0019] Take the union of the initial demand label and the corresponding label after statistics as the user demand label;

[0020] The calculation process of the demand value corresponding to the user demand label includes:

[0021] By formula:

[0022]

[0023] Calculate the demand value of the i-th user demand label ;

[0024] in, is the initial demand label judgment function. When the i-th user demand label exists in the initial demand label, ;otherwise, ; is the number of the i-th user requirement label that exists in the corresponding label after statistics, is the number of times the i-th user demand label appears in the user's historical teaching information; p and are all parameter adjustment coefficients, satisfying p< <1.

[0025] By adopting the above technical solution, a calculation process for the demand value of the user demand label is given. By calculating the demand value of each user demand label obtained, the user's subjective demand, the user's historical teaching information and the degree of demand for teaching test information can be determined, thereby facilitating the subsequent targeted matching process.

[0026] Optionally, the process of extracting teaching content by the teaching content label extraction unit includes:

[0027] Obtain the curve V(t) of the change of playback volume with video time bar in the historical playback data of teaching content, derive the curve V(t) of the change of playback volume with video time bar, and obtain the curve of the change rate of playback volume with video time bar ;

[0028] According to V(t) and Determine the extraction of teaching content.

[0029] By adopting the above technical solution, according to V(t) and To determine the extraction of teaching content, not only the time period with higher playback popularity value can be used as the extracted content, but also the peak time period with obvious difference in playback popularity can be used as the extracted content, thereby realizing accurate division and extraction of course content.

[0030] Optionally, according to V(t) and The extraction process of determining teaching content includes:

[0031] V(t) and To compare, The corresponding interval is used as the first extraction interval;

[0032] Will Compare with the preset fixed values ​​k1 and -k1, The solution of is taken as the first solution set, The solution of is taken as the second solution set, and any solution in the first solution set to any solution in the adjacent second solution set with a value greater than the solution in the first solution set is taken as the second extraction interval;

[0033] A union of the first extraction interval and the second extraction interval is obtained, the teaching content corresponding to the union is used as the extracted content, and a label of the extracted content is obtained.

[0034] By adopting the above technical solution, the union of the first extraction interval and the second extraction interval is taken, the teaching content corresponding to the union is used as the extracted content, and the label of the extracted content is obtained. It should be noted that the acquisition of the extracted content label can be obtained through worker annotation or through big data analysis of the extracted content. Through the label of the extracted content, the content of the time interval can be accurately judged, which is convenient for the subsequent matching process.

[0035] Optionally, the calculation process of extracting the content recommendation value includes:

[0036] By formula:

[0037]

[0038] Calculate and obtain the recommendation value R of each extracted content;

[0039] Where t1 is the initial timestamp of each extracted content, and t2 is the end timestamp of each extraction. is the preset value, G is the evaluation value of the current extracted content, is the average evaluation value of all content on the platform, is the adjustment factor.

[0040] By adopting the above technical solution, the recommendation status of the extracted content can be accurately evaluated by extracting the relative degree of the content playback volume to the overall degree of the curve of the video time bar and the relevant evaluation data, which is convenient for subsequent corresponding matching according to the recommendation value of each extracted content, and thus can better select more accurate and suitable content for recommendation according to user needs, thereby improving the accuracy of recommendations and the adaptability to different users.

[0041] Optionally, the process of recommendation by the matching recommendation module includes:

[0042] The teaching content in the teaching platform is preliminarily screened according to the N tags with the highest user demand values, and then the screened teaching content is matched and calculated with the user demand tags and the corresponding demand values ​​to obtain the matching value, and the best recommended content for the user is obtained according to the size of the matching value.

[0043] By adopting the above technical solution, through preliminary screening, the amount of calculation in the subsequent matching process can be greatly reduced. Then, through the calculation of the matching value, the process of accurately matching user needs with teaching content can be achieved more accurately, thereby improving the accuracy of recommendations and the adaptability to different users.

[0044] Optionally, the calculation process of the matching value includes:

[0045] By formula:

[0046]

[0047] Calculate the matching value between the i-th user and the k-th teaching content after screening ;

[0048] Select the teaching content corresponding to the maximum matching value as the best recommended content and recommend it to the user;

[0049] in, is the number of labels that overlap between the i-th user and the k-th teaching content after screening, j=1, 2, …, , is the demand value of the jth tag of the ith user, is the recommended value of the jth label of the kth teaching content.

[0050] By adopting the above technical solution, the more the user's needs match a certain teaching content, the more overlapping labels there are, the greater the demand value and the recommendation value, and the higher the final matching value. Therefore, the teaching content corresponding to the maximum matching value is selected as the best recommended content to recommend to the user, and then the content that best matches the user's needs can be recommended to the user, which improves the accuracy of the recommendation and the adaptability to different users.

[0051] In summary, the present application includes at least one of the following beneficial technical effects:

[0052] 1. The present invention performs user demand analysis based on the user's input information, historical teaching information and teaching test information, obtains user demand labels, and also determines demand values ​​corresponding to different demand labels. The demand value is not only determined based on the user's input information to meet the user's subjective needs, but also objectively analyzed based on the historical teaching information and teaching test information to meet the user's objective needs. Therefore, the obtained demand value can accurately reflect the user's real objective demand status for the teaching course. In addition, the present invention extracts the teaching content, obtains the extracted content label, and also obtains the corresponding recommendation value based on the historical playback data and user evaluation data of the extracted content. Compared with the direct label matching method, the label and the corresponding recommendation value matching method can better select more accurate and suitable content for recommendation based on the user's needs, thereby improving the accuracy of the recommendation and the adaptability to different users.

[0053] 2. The present invention obtains the user's wrong question information and its corresponding labels in the teaching test information, counts the corresponding labels of the wrong question information, compares the counted corresponding labels with the user's historical teaching information, obtains the comparison results, and then determines the user's demand labels and their corresponding demand values ​​according to the initial demand labels, the counted corresponding labels and the comparison results, and then can accurately reflect the user's real objective demand status for the teaching course according to the obtained demand values, thereby improving the accuracy of recommendations and the adaptability to different users.

[0054] 3. The present invention preliminarily screens the teaching content in the teaching platform according to the N tags with the highest user demand values, and then matches and calculates the screened teaching content with the user demand tags and the corresponding demand values ​​to obtain matching values, and obtains the best recommended content for the user according to the size of the matching value. The preliminary screening method can greatly reduce the amount of calculation in the subsequent matching process, and then through the calculation process of the matching value, the process of accurately matching user needs with teaching content can be achieved more accurately, thereby improving the accuracy of recommendations and the adaptability to different users.

[0055] 4. The present invention uses the matching value calculation process. When the user demand is more matched with a certain teaching content, the more overlapping labels there are, the greater the demand value and the recommendation value are, and the higher the final matching value is. Therefore, the teaching content corresponding to the maximum matching value is selected as the best recommended content to recommend to the user, and then the content that best matches the user's demand can be recommended to the user, which improves the accuracy of the recommendation and the adaptability to different users. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a logical block diagram of a personalized course intelligent recommendation management platform based on big data. DETAILED DESCRIPTION

[0057] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings.

[0058] In the description of this specification, the description with reference to the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0059] The present application embodiment discloses a personalized course intelligent recommendation management platform based on big data, Figure 1, including a user data collection end, a user demand analysis end, a teaching content label extraction unit and a matching recommendation module, wherein the user data collection end is used to obtain the user's input information, historical teaching information and teaching test information; the user demand analysis end performs user demand analysis based on the user's input information, historical teaching information and teaching test information, and obtains user demand labels and demand values ​​corresponding to different demand labels; the teaching content label extraction unit extracts the teaching content, obtains the label of the extracted content, and obtains the corresponding recommendation value based on the historical playback data and user evaluation data of the extracted content; the user demand label and the demand values ​​corresponding to different demand labels are matched with the label of the teaching content and the corresponding recommendation value through the matching recommendation module to obtain the user's best recommended content. Compared with the existing personalized course intelligent recommendation system, this embodiment is based on the user's input The user demand analysis is performed based on the user demand label, historical teaching information and teaching test information. While obtaining the user demand label, the demand values ​​corresponding to different demand labels are also determined. The demand value is not only determined based on the user's input information to meet the user's subjective needs, but also objectively analyzed based on the historical teaching information and teaching test information to meet the user's objective needs. Therefore, the obtained demand value can accurately reflect the user's real objective demand status for the teaching course. In addition, this embodiment extracts the teaching content, obtains the extracted content label, and also obtains the corresponding recommendation value based on the historical playback data and user evaluation data of the extracted content. Compared with the direct label matching method, the label and the corresponding recommendation value matching method can better select more accurate and suitable content for recommendation based on the user's needs, thereby improving the accuracy of the recommendation and the adaptability of different users.

[0060] In addition, the process of user demand analysis includes: labeling the user's input information based on big data to obtain initial demand labels. This process is based on AI comparison of the content corresponding to different labels in the big data. When the comparison overlap reaches a preset value, it is judged that the user's demand content is consistent with the label, so the label is used as the initial demand label. The specific operation process is the existing technology and is not further described here; then the user's wrong question information and its corresponding labels in the teaching test information are obtained, the corresponding labels of the wrong question information are counted, and the corresponding labels after statistics are compared with the user's historical teaching information to obtain the comparison results, wherein the acquisition of the corresponding labels of the wrong question information can be obtained directly according to the corresponding wrong question information by pre-marking, and the wrong question information can also be analyzed based on big data, and the corresponding labels are obtained according to the analysis results. There is no restriction here, and then the user demand label and its corresponding demand value are determined according to the initial demand label, the corresponding label after statistics and the comparison result, so that the user's real objective demand state for the teaching course can be accurately reflected according to the obtained demand value, thereby improving the accuracy of the recommendation and the adaptability of different users.

[0061] In one embodiment, the process of determining the user demand label and its corresponding demand value includes: taking the union of the initial demand label and the corresponding label after statistics as the user demand label; the process of calculating the demand value corresponding to the user demand label includes:

[0062] By formula:

[0063]

[0064] Calculate the demand value of the i-th user demand label ;

[0065] in, is the initial demand label judgment function. When the i-th user demand label exists in the initial demand label, ;otherwise, ; is the number of the i-th user requirement label that exists in the corresponding label after statistics, is the number of times the i-th user demand label appears in the user's historical teaching information; p and are all parameter adjustment coefficients, satisfying p< <1, parameter adjustment coefficient p and The setting is selected based on experience according to the numerical range in the simulation data to adjust the degree of influence of the number of user demand labels in the corresponding labels after statistics and the number of occurrences of the user's historical teaching information on the demand value. Therefore, through the calculation process of the demand value of the user demand label, the user's subjective needs, the user's historical teaching information and the degree of demand for teaching test information can be determined, which is convenient for the subsequent targeted matching process.

[0066] The process of extracting teaching content by the teaching content tag extraction unit includes: obtaining a curve V(t) showing a change in the playback volume versus the video time bar in the historical playback data of the teaching content. It should be noted that the statistical method for the video playback volume is to use the total playback volume as a method, and also to count the number of playbacks at each time point for different clips, that is, the curve V(t) showing a change in the playback volume versus the video time bar; deriving the curve V(t) showing a change in the playback volume versus the video time bar to obtain a curve showing a change rate of the playback volume versus the video time bar ; According to V(t) and To determine the extraction of teaching content, not only the time period with higher playback popularity value can be used as the extracted content, but also the peak time period with obvious difference in playback popularity can be used as the extracted content, thereby realizing accurate division and extraction of course content.

[0067] Among them, according to V(t) and The extraction process of determining teaching content includes: V(t) and For comparison, when , it means that the corresponding content popularity is higher than the average state. The corresponding interval is taken as the first extraction interval, so the content corresponding to the first extraction interval is the key content of the teaching content; Compare with the preset fixed values ​​k1 and -k1. The value of k1 is selected and set according to empirical data. The solution of is taken as the first solution set, The solution is taken as the second solution set, and any solution in the first solution set to any solution in the adjacent second solution set that is greater than its value is taken as the second extraction interval. Obviously, the content corresponding to the second extraction interval is higher than the average state of the video, so the content corresponding to the second extraction interval is also the key content of the teaching content; take the union of the first extraction interval and the second extraction interval, take the teaching content corresponding to the union as the extracted content, and obtain the label of the extracted content. It should be noted that the acquisition of the extracted content label can be obtained through worker annotation or through big data analysis of the extracted content. Through the label of the extracted content, the content of the time interval can be accurately judged, which is convenient for the subsequent matching process.

[0068] In one embodiment, a calculation process for extracting content recommendation value is provided, including: using the formula:

[0069]

[0070] Calculate and obtain the recommended value R of each extracted content; where t1 is the initial timestamp of each extracted content, and t2 is the end timestamp of each extraction. is a preset value, which is set according to the average value corresponding to the similar type, G is the evaluation value of the current extracted content, is the average evaluation value of all the content on the platform. It should be noted that the evaluation value is obtained using existing technology and is not limited here; It is an adjustment coefficient, which is selected and set based on empirical data. Therefore, by extracting the relative degree of the overall degree of the curve of the content playback volume changing with the video time bar and the relevant evaluation data, an accurate evaluation of the recommendation status of the extracted content can be achieved, which is convenient for subsequent matching according to the recommendation value of each extracted content. Therefore, it is able to better select more accurate and suitable content for recommendation based on user needs, thereby improving the accuracy of recommendations and the adaptability to different users.

[0071] In addition, the process of recommendation by the matching recommendation module includes: preliminary screening of the teaching content in the teaching platform according to the N tags with the highest user demand values. It should be noted that the value of N is set by the administrator of the course teaching platform, and is mainly selected according to the richness of the content of the course teaching platform. When the resource level in the course teaching platform is high, the number of N can be appropriately increased to reduce the problem of omissions in the screening process. When the resource level in the course teaching platform is slightly lower, the number of N can be appropriately reduced to improve the screening speed while meeting the needs; then the screened teaching content is matched and calculated with the user demand tag and the corresponding demand value to obtain the matching value, and the best recommended content for the user is obtained according to the size of the matching value. In the above scheme, the preliminary screening method can greatly reduce the amount of calculation in the subsequent matching process one by one, and then through the calculation process of the matching value, the process of accurately matching user needs with teaching content can be more accurately realized, thereby improving the accuracy of the recommendation and the adaptability of different users.

[0072] In one embodiment, the calculation process of the matching value includes: using the formula:

[0073]

[0074] Calculate the matching value between the i-th user and the k-th teaching content after screening ;

[0075] Select the teaching content corresponding to the maximum matching value as the best recommended content and recommend it to the user;

[0076] in, is the number of labels that overlap between the i-th user and the k-th teaching content after screening, j=1, 2, …, , is the demand value of the jth tag of the ith user, is the recommendation value of the jth label of the kth teaching content. It should be noted that the recommendation values ​​of all labels corresponding to each teaching content are equal to the recommendation value of the teaching content. Therefore, the more the user's needs match a certain teaching content, the more overlapping labels there are, the greater the demand value and the recommendation value, and the higher the final matching value. Therefore, the teaching content corresponding to the maximum matching value is selected as the best recommended content to recommend to the user, and then the content that best matches the user's needs can be recommended to the user, which not only improves the accuracy of the recommendation and the adaptability to different users.

[0077] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A personalized course intelligent recommendation management platform based on big data, characterized by: include: User data collection terminal, used to obtain user input information, historical teaching information and teaching test information; The user demand analysis terminal is used to analyze user demand based on user input information, historical teaching information and teaching test information, and obtain user demand tags and demand values ​​corresponding to different demand tags; A teaching content label extraction unit is used to extract teaching content, obtain labels of the extracted content, and obtain corresponding recommendation values ​​according to historical playback data and user evaluation data of the extracted content; The matching recommendation module is used to match the user's demand labels and the demand values ​​corresponding to different demand labels with the labels of the teaching content and the corresponding recommendation values ​​to obtain the best recommended content for the user.

2. According to the big data-based personalized course intelligent recommendation management platform of claim 1, it is characterized by: The process of user needs analysis includes: Label the user's input information based on big data to obtain the initial demand labels; Obtain the user's wrong question information and its corresponding labels in the teaching test information, count the corresponding labels of the wrong question information, compare the corresponding labels after counting with the user's historical teaching information, and obtain the comparison results; Determine the user demand label and its corresponding demand value based on the initial demand label, the corresponding label after statistics and the comparison result.

3. According to claim 2, a personalized course intelligent recommendation management platform based on big data is characterized in that: The process of determining the user requirement label and its corresponding requirement value includes: Take the union of the initial demand label and the corresponding label after statistics as the user demand label; The calculation process of the demand value corresponding to the user demand label includes: By formula: Calculate the demand value of the i-th user demand label ; in, is the initial demand label judgment function. When the i-th user demand label exists in the initial demand label, ;otherwise, ; is the number of the i-th user requirement label that exists in the corresponding label after statistics, is the number of times the i-th user demand label appears in the user's historical teaching information; p and are all parameter adjustment coefficients, satisfying p< <1.

4. According to the big data-based personalized course intelligent recommendation management platform of claim 3, it is characterized in that: The process of extracting teaching content by the teaching content label extraction unit includes: Obtain the curve V(t) of the change of playback volume with video time bar in the historical playback data of teaching content, derive the curve V(t) of the change of playback volume with video time bar, and obtain the curve of the change rate of playback volume with video time bar ; According to V(t) and Determine the extraction of teaching content.

5. According to claim 4, a personalized course intelligent recommendation management platform based on big data is characterized in that: According to V(t) and The extraction process of determining teaching content includes: V(t) and To compare, The corresponding interval is used as the first extraction interval; Will Compare with the preset fixed values ​​k1 and -k1, The solution of is taken as the first solution set, The solution of is taken as the second solution set, and any solution in the first solution set to any solution in the adjacent second solution set with a value greater than the solution value is taken as the second extraction interval; A union of the first extraction interval and the second extraction interval is obtained, the teaching content corresponding to the union is used as the extracted content, and a label of the extracted content is obtained.

6. The personalized course intelligent recommendation management platform based on big data according to claim 5 is characterized in that: The calculation process of extracting the content recommendation value includes: By formula: Calculate and obtain the recommendation value R of each extracted content; Where t1 is the initial timestamp of each extracted content, and t2 is the end timestamp of each extraction. is the preset value, G is the evaluation value of the current extracted content, is the average evaluation value of all content on the platform, is the adjustment factor.

7. The personalized course intelligent recommendation management platform based on big data according to claim 6 is characterized in that: The process of recommendation by the matching recommendation module includes: The teaching content in the teaching platform is preliminarily screened according to the N tags with the highest user demand values, and then the screened teaching content is matched and calculated with the user demand tags and the corresponding demand values ​​to obtain the matching value, and the best recommended content for the user is obtained according to the size of the matching value.

8. The personalized course intelligent recommendation management platform based on big data according to claim 7 is characterized in that: The calculation process of the matching value includes: By formula: Calculate the matching value between the i-th user and the k-th teaching content after screening ; Select the teaching content corresponding to the maximum matching value as the best recommended content and recommend it to the user; in, is the number of labels that overlap between the i-th user and the k-th teaching content after screening, j=1, 2, …, , is the demand value of the jth tag of the ith user, is the recommended value of the jth label of the kth teaching content.

Images