CN112163161B - Recommendation method and system for college library, readable storage medium and electronic equipment - Google Patents
Recommendation method and system for college library, readable storage medium and electronic equipment Download PDFInfo
- Publication number
- CN112163161B CN112163161B CN202011096851.2A CN202011096851A CN112163161B CN 112163161 B CN112163161 B CN 112163161B CN 202011096851 A CN202011096851 A CN 202011096851A CN 112163161 B CN112163161 B CN 112163161B
- Authority
- CN
- China
- Prior art keywords
- students
- books
- preferences
- feature
- attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Strategic Management (AREA)
- Educational Administration (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Educational Technology (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a recommendation method and a recommendation system for a library of colleges and universities, a readable storage medium and electronic equipment, wherein the recommendation method for the library of colleges and universities comprises the following steps: capturing individual preferences of students on books of different classes, and capturing professional preferences of students on books of different classes; and establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, so as to recommend the books matched with the preferences of the students. The recommendation method, the recommendation system, the readable storage medium and the electronic equipment for the library of colleges and universities provide accurate book recommendation by a method of combining the relevance between the student subjects and the book types and the individual diversity, so that not only can the library resources be fully utilized, but also the learning achievement of students can be improved.
Description
Technical Field
The invention belongs to the technical field of neural networks, relates to a recommendation method and a recommendation system, and particularly relates to a recommendation method and a recommendation system for a library of colleges and universities, a readable storage medium and electronic equipment.
Background
College libraries have rich library resources, and students can acquire information needed by themselves in the library. However, the large amount of resources also makes it difficult for students to accurately and efficiently find books that they are interested in. Therefore, the recommendation system is worthy of extensive research in libraries of colleges and universities as an effective solution for overcoming information overload. In addition, the circulation records of the books accumulated by the library provide rich context perception information (such as student specialties and book categories) for students, and can be used for describing the behavior characteristics of the students. Providing accurate book recommendation not only can make full use of library resources, but also is helpful for improving the learning achievement of students [1].
Unlike book recommendations for e-commerce websites, the book recommendations for libraries are mainly directed to students who have distinct distinguishing attributes: professional. Therefore, students can borrow a large number of professional lesson books according to the requirements of professional lessons. Library book recommendations have the following three features:
(1) First, students have strong professional pertinence in book borrowing. We counted the general case of book borrowing for students from different colleges (or specialties), as shown in fig. 1. It is clear that most students are closely related to their professions in the category of books they borrow the most. The second or third most borrowed category appears to be professional independent. Therefore, we consider that the preferences of students are composed of two parts, personal interests and professional preferences. For example, students in computer science specialties like physics and literature in addition to computer-like books, and their borrowing history may include TP, I2, and G8.TP relates to his discipline, while I2 and G8 relate to his personal interests. Empirically, personal interests and professional preferences may cross, but have little impact on our research. It is readily apparent that different categories are of different importance to different disciplines. Obviously, in this case, the student should be recommended books that are closely related to the course of the school. Therefore, recommendation algorithms need to be used to learn the correlations between book categories and disciplines.
(2) Second, professional preferences vary with the educational phase. The primary education of Chinese university is generally four years, and the professional ability of students is required to be improved along with the improvement of grade. More professional lessons in senior levels are available than basic lessons and general lessons in senior levels. For example, students of computer science specialities learn basic courses, such as programming, at a lower level, but they host specialized core courses, such as compilation principles, at a higher level. Thus, students in the same profession may have slightly different professional preferences at different grades.
(3) Finally, library recommendations present strong sparsity, as most students borrow only a small number of books each year (compared to the collection), which presents a huge challenge to the recommendation algorithm.
Conventional library book recommendation methods are generally based on association rules that specify a user's interest in a given book type. Although mining the association between student disciplines and book types has facilitated library book recommendations, these association rule-based recommendation methods fall into coarse-grained recommendations due to neglecting individual diversity. The most well-known method in recommendation systems is collaborative filtering, which makes recommendations for users from similar users or similar items based on the users' historical behavior. However, these methods are not suitable for application to large-scale, sparse, real datasets. SCWMF is the first and most well-known method of generating library book recommendations on a large real-world dataset, based on context-aware matrix factorization (CMF). The SCWMF effectively models hidden vectors of users and articles by combining learning scores of students and meta information of books (e.g., book categories, etc.), thereby improving recommendation performance. While CMF-based methods employ more ancillary information to effectively represent hidden vectors of users and items, these methods still only consider interactions between two entities (i.e., users and items), ignoring more interactions between contextual features. In fact, book recommendations require more feature interactions to be considered. To provide personalized library book recommendations, we need to accurately shape their interests from the students' historical behavior, with little work related to library book recommendations. Embedding is a very efficient method of feature representation, which projects high-dimensional sparse vectors into low-dimensional hidden vectors. Therefore, to obtain more effective recommendations, models must be designed for better user and item expression. Furthermore, the embedding dimension may contain richer semantics. On the basis, the ONCF obtains a characteristic diagram by utilizing an outer product on an embedded layer, and then extracts a high-order signal by using a neural network. We believe that the disadvantage of this work is that one feature map cannot capture the higher order interaction signals between more than three features and therefore this approach is not sufficient to learn the correlation between multiple features and extract the key information.
Therefore, how to provide a recommendation method, a recommendation system, a readable storage medium and an electronic device for a library in colleges and universities to solve the defects that the prior art ignores more interaction between context features, is not enough to learn the association among the features, cannot extract key information, and causes inaccurate book recommendation, and the like, has become a technical problem to be solved by technical personnel in the field.
Disclosure of Invention
In view of the above shortcomings of the prior art, an object of the present invention is to provide a recommendation method, a recommendation system, a readable storage medium and an electronic device for libraries in colleges and universities, which are used to solve the problems of inaccurate book recommendation caused by the prior art that more interactions between contextual features are ignored, the association between multiple features is not sufficiently learned, and key information cannot be extracted.
To achieve the above and other related objects, an aspect of the present invention provides a recommendation method for a library of colleges and universities, including: the method comprises the following steps of capturing individual preferences of students on books of different categories, and capturing professional preferences of the students on the books of different categories; and establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, so as to recommend the books matched with the preferences of the students for the students.
In an embodiment of the present invention, the step of capturing personalized preferences of individual students for different categories of books includes: receiving discrete attribute data related to a student borrowing a book; mapping the discrete attribute data into dense attribute data related to the student borrowing books; and establishing an interaction model of each pair of attribute features in the dense attribute data.
In an embodiment of the present invention, the step of capturing personalized preferences of individual students for different categories of books further includes: receiving historical borrowing records of students; and establishing the weighted preference of all the borrowing categories of the students according to the historical borrowing records of the students.
In an embodiment of the present invention, the step of capturing personalized preferences of individual students for different categories of books further includes: the weighted preference of all the borrowing categories of the students is interacted with the attribute characteristics to form an interaction result; and combining the interaction result with the interaction model of each pair of attribute characteristics to obtain the personalized preference of the individual student to the books of different classes.
In an embodiment of the present invention, the step of capturing the professional preferences of the student for books of different categories includes: selecting an attribute feature related to the college from a plurality of attribute features; forming a feature interaction graph of attribute features related to the college; convolving the feature interaction graph to form a new multi-channel feature graph; the new multi-channel feature map is used for representing the correlation between every two attribute features; the correlation between two of the attribute features is mapped to represent a professional preference of a student for a particular specialty at a particular school.
In an embodiment of the present invention, the step of convolving the feature interaction map to form a new multi-channel feature map includes: inputting the feature interactive map of the attribute features related to the college into a convolution layer of a convolution neural network, using a two-dimensional convolution kernel, adding a bias, and performing nonlinear transformation by using an activation function to form a new multi-channel feature map.
In an embodiment of the present invention, the preference model of the student = (1- λ) × personalized preferences of individual students for different classes of books + λ × professional preferences of a student for a specific specialty at a specific school date; λ is a hyper-parameter controlling two interest weights.
In another aspect, the present invention provides a recommendation system for libraries in colleges and universities, including: the first capturing module is used for capturing the personalized preferences of students on books of different categories; the second capturing module is used for capturing professional preferences of students on books of different categories; and the processing module is used for establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories so as to recommend the books matched with the preferences of the students.
Yet another aspect of the present invention provides a readable storage medium having stored thereon a computer program that, when executed by a processor, implements the recommendation method for a library of colleges and universities.
A final aspect of the present invention provides an electronic device, comprising: a processor and a memory; the memory is used for storing a computer program, and the processor is used for executing the computer program stored by the memory so as to enable the electronic equipment to execute the recommendation method of the library of colleges and universities.
As described above, the recommendation method, system, readable storage medium and electronic device for libraries in colleges and universities according to the present invention have the following advantages:
the recommendation method, the recommendation system, the readable storage medium and the electronic equipment for the library of colleges and universities provide accurate book recommendation by a method of combining the relevance between the student subjects and the book types and the individual diversity, so that not only can the library resources be fully utilized, but also the learning achievement of students can be improved.
Drawings
Fig. 1 is a flowchart illustrating a recommendation method for libraries in colleges and universities according to an embodiment of the present invention.
Fig. 2 is a schematic diagram illustrating a process of establishing a preference model of a student according to the present invention.
FIG. 3 shows the effect of professional preference on the A university data set.
FIG. 4 is a schematic diagram illustrating a schematic structure of a recommendation system for libraries in colleges and universities according to an embodiment of the present invention.
Description of the element reference numerals
4. Recommendation system for libraries in colleges and universities
41. First capture module
42. Second capture module
43. Processing module
S11 to S12 steps
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
Example one
The embodiment provides a recommendation method for a library of colleges and universities, which comprises the following steps:
capturing individual preferences of students on books of different classes, and capturing professional preferences of students on books of different classes;
and establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, so as to recommend the books matched with the preferences of the students for the students.
The following sets of diagrams to describe in detail the recommendation method for libraries in colleges and universities provided by the present embodiment. Please refer to fig. 1 and fig. 2, which are a schematic flow chart of a recommendation method for a library in colleges and universities in an embodiment and a schematic process diagram of a preference model establishment for students, respectively. As shown in fig. 1, the recommendation method for libraries in colleges and universities specifically includes the following steps:
s11, capturing the individual preference of the students to the books of different classes, and capturing the professional preference of the students to the books of different classes.
In the embodiment, the personalized preferences of the students on the books of different categories and the professional preferences of the students on the books of different categories are captured through the convolutional neural network.
Wherein the step of capturing individual student's personalized preferences for different classes of books further comprises:
and receiving the historical borrowing records of the students. In thatIn this embodiment, the student history lending record x h ∈R c And c represents the total number of the categories according to the borrowing times of each category. For example, the history borrowing record is x h ＝[1,2,0,0，…，1]The number of the borrowed books in the first category of the student is 1, the number of the borrowed books in the second category of the student is 2, and the like.
And establishing the weighted preference of all the borrowing categories of the students according to the historical borrowing records of the students.
Since the student's circulation history is very specific, it should be weighted more heavily if the book is more relevant to the student's subject. Because students are more likely to borrow those books in the future. Although similarity calculation between categories may solve this problem, it may ignore the user's inherent preferences.
For example, according to the mesographic classification, both TP and TN are related to Computer Science and technology (CS), but for CS students who may be more interested in Computer network communications, TN books are heavily borrowed. To accurately evaluate the individual preferences of students for different categories, category similarity must be combined with the student's borrowing history. Thus, attention mechanisms may be effective in solving this problem. Professional targeted and individualized differences are addressed by multi-layered perceptrons (MLPs), which are defined as:
a i ＝h T tanh(Wk i +Uv 2 )
wherein v is 2 ∈R d Embedded vector, k, representing student college i An embedded vector representing class i, W and U represent weight matrices for two different vectors, and h maps the hidden state to an attention score. The weight of each person can then be found by the Softmax function:
where C represents a set of categories and t i A normalized representation of the number of borrowings for each category, with a weighted preference for all categories:
the step of capturing individual student's personalized preferences for different classes of books comprises:
first, discrete attribute data relating to a student borrowing a book is received. In this embodiment, the discrete attribute data associated with the student borrowing the book is used to describe a particular student context and book attributes, including unique heat coding features (e.g., student ID) and real-valued features (e.g., student history borrowing number).
As shown in fig. 2, discrete attribute data relating to book borrowing by students is received, specifically 5 fields of sparse data.
The discrete attribute data is then mapped into dense attribute data relating to the student borrowing the book.
Specifically, the discrete attribute data is mapped into dense attribute data related to the student borrowing books through embedded table query, and the dense attribute data related to the student borrowing books is expressed as d-dimensional vectors.
As shown in fig. 2, due to the sparsity of the input vector, non-0 features need to be considered, and v1 (representing the student ID in this embodiment), v2 (representing the specialty in this embodiment), v3 (representing the specific school date in this embodiment), v4 (representing the personalized preference of the student in this embodiment), v5 (representing the book category in this embodiment), and v6 (representing the book name of the book in this embodiment) are found through the embedded layer.
And then, establishing an interaction model of each pair of attribute features in the dense attribute data.
Specifically, each pair of feature interactions in v1 (representing student ID in this embodiment), v2 (representing professional in this embodiment), v3 (representing specific school date in this embodiment), v4 (representing personalized preference of student in this embodiment), and v5 (representing book category in this embodiment) is modeled. For a given multi-domain feature x ∈ R n . The modeling method specifically comprises the following steps:
wherein, ω is 0 Representing a global deviation, ω i A weight value representing the ith variable is shown,<v i ,v j >representing the inner product of two vectors, v i ∈R d Is the embedded vector of the ith feature and d represents the size of the vector.
The weighted preference of all the borrowing categories of the students is interacted with the attribute characteristics to form an interaction result
And combining the interaction result with the interaction model of each pair of attribute characteristics to obtain the personalized preference of the individual student to the books of different classes.
Specifically, the individual students have personalized preferences for different classes of books as
The step of capturing the professional preferences of the students for different categories of books comprises the following steps:
first, an attribute feature related to the college is selected from a plurality of attribute features.
In the present embodiment, the attribute features related to the college include v2 (representing profession in the present embodiment), v3 (representing specific school date in the present embodiment), and v6 (representing book title of book in the present embodiment).
Then, a feature interaction graph of the attribute features related to the college is formed.
In the present embodiment, M is used i,j ∈R d×d Representation feature v i And v j And (5) feature interaction. Although the outer product between the embedded vectors can obtain a better representation, the interaction of features that are significantly independent of the target vector makes the neural network redundant, extending the training time for each iteration. Since the object of this embodiment is to understand the relationship between student majors and student term and title,i, j is limited to {2,3,6}. Then, a "three-channel" stacked feature map can be obtained, just like an RGB image, and the interaction model for each pair of attribute features can be expressed as:
χ＝[M 2,3 ，M 2,6 ，M 3,6 ]
the interaction model carries higher order signals, e.g. M 2,3 And M 2,6 The interaction between v2 and v6 can be regarded as a higher order interaction between v3 and v 6.
Then, carrying out convolution on the feature interaction graph to form a new multi-channel feature graph; the new multi-channel feature map is used for representing the correlation between every two attribute features.
Specifically, a feature interaction diagram of the attribute features related to the college is input into a convolution layer of a convolution neural network, a two-dimensional convolution kernel is used, after bias is added, nonlinear transformation is carried out by utilizing an activation function, and a new multichannel feature diagram is formed.
The new multi-channel signature is as follows:
χ l+1 ＝ReLU(b l+1 +χ l ⊙K l+1 )
wherein, b l+1 Denotes the deviation of the l +1 layer, K l+1 A convolution kernel indicating an l +1 layer, ", indicates a convolution operation. Assuming the embedding dimension d is 128, the size of the stacked feature interaction graph is 128 x 3. We select a convolution kernel of size 2*2 and set the step size to 2*2 at each layer and then get a feature interaction map with the multichannel height and width half that of the previous layer.
The correlation between two of the attribute features is mapped to a professional preference that represents a student's specific specialty for a particular school period. The professional preference of the student for a specific professional in a specific school period is expressed by a real-value scalar, and the specific conditions are as follows:
and S12, establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, and recommending the books matched with the preferences of the students.
Specifically, the preference model of the student = (1- λ) × personalized preferences of individual students for different classes of books + λ × professional preferences of the student's specific specialty in a specific school date; lambda is a hyper-parameter controlling two interest weights,
namely that
Wherein, lambda belongs to [0,1 ]]Is a hyper-parameter that controls two interest weights.
In this example, n = {5,10,20} is selected to compare Top-n recommended performance on our dataset. Table 2,3,4 shows the performance of our method compared to four baselines on the primary test criteria. Obviously, it can be seen that for each value of n, the proposed recommendation algorithm of the present embodiment is superior to other algorithms. The results show that feature interactions generated by outer products can extract a richer signal than feature interactions based on element dot products (NFM). In addition, the embedded expression of each characteristic can be obtained more effectively by combining personal interest and professional preference compared with the result of the SCWMF, so that the recommended performance is improved.
Table 1: comparison of Performance of datasets at university A
Table 2: comparison of Performance of datasets at university B
Table 3: comparison of Performance of datasets at university C
From the objective function of the student's preference model ACFM, the influence of two parts of the lambda control model on the prediction result can be seen. It was found that considering personal interests only (λ = 0) generally performed better than considering professional preferences only (λ = 1.0), as shown in figure 3 for the effect of university a dataset professional preferences. In addition, the performance is better when the value of lambda is controlled between 0.4 and 0.5, which shows that personal interest plays a more important role in library book recommendation.
In this example, the effectiveness of our algorithm was evaluated by off-line experiments on three real-scene datasets containing library circulation records of three different universities for four consecutive years, with statistical data as shown in table 4. Our dataset includes graduate and non-graduate bibliographic information from most colleges. This study uses borrowing as implicit feedback. In order to fully understand the preference of students, the students with less than 10 books per school period are screened, and the books with less than 2 books are filtered. We selected six context-dependent properties from the data set for our experiments. The user context consists of student ID, student college and student school period. The item attributes include book ID, book category, and book name, where the book name is represented by a word2vec vector.
Table 4: circulation record statistics for three college libraries
The recommendation method for the library of colleges and universities, which is disclosed by the embodiment, provides accurate book recommendation by a method of combining the relevance between the student subjects and the book types and the individual diversity, so that not only can the library resources be fully utilized, but also the learning achievement of students can be improved.
The present embodiment provides a readable storage medium (also referred to as a computer-readable storage medium) having stored thereon a computer program which, when executed by a processor, implements the recommendation method for a library of colleges and universities as recited in any one of claims 1 to 7.
One of ordinary skill in the art will appreciate that the computer-readable storage medium is: all or a portion of the steps of implementing the above-described method embodiments may be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
Example two
The embodiment provides a recommendation system for libraries in colleges and universities, which comprises:
the first capturing module is used for capturing the personalized preferences of students on books of different categories;
the second capturing module is used for capturing professional preferences of students on books of different categories;
and the processing module is used for establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories so as to recommend the books matched with the preferences of the students.
The recommendation system for libraries in colleges and universities provided by the present embodiment will be described in detail with reference to the drawings. The recommendation system 4 of the library of colleges and universities of the embodiment includes a first capturing module 41, a second capturing module 42 and a processing module 43.
The first capturing module 41 is used for capturing personalized preferences of students for different classes of books.
Specifically, the first capturing module 41 receives discrete attribute data related to book borrowing by students; mapping the discrete attribute data into dense attribute data related to the student borrowing books; establishing an interaction model of each pair of attribute features in the dense attribute data;
the first capturing module 41 is further used for receiving the historical borrowing records of students; establishing the weighting preference of all borrowing categories of the students according to the historical borrowing records of the students; the weighted preference of all the borrowing categories of the students is interacted with the attribute characteristics to form an interaction result; and combining the interaction result with the interaction model of each pair of attribute characteristics to obtain the personalized preference of the individual student to the books of different classes.
The second capture module 42 is used for capturing professional preferences of students for books of different categories.
Specifically, the second capturing module 42 selects an attribute feature related to the college from a plurality of attribute features; forming a feature interaction graph of attribute features related to the college; convolving the feature interaction graph to form a new multi-channel feature graph; the new multi-channel feature map is used for representing the correlation between every two attribute features; the correlation between two of the attribute features is mapped to a professional preference that represents a student's specific specialty for a particular school period.
When the second capture module 42 convolves the feature interaction map to form a new multi-channel feature map, the feature interaction map of the attribute features related to the college is input to a convolution layer of a convolutional neural network, a two-dimensional convolution kernel is used, and after a bias is added, nonlinear transformation is performed by using an activation function to form the new multi-channel feature map.
In this embodiment, the preference model of the student = (1- λ) × personalized preferences of individual students for different classes of books + λ × professional preferences of the student's specific specialty in a specific school date; λ is a hyper-parameter controlling two interest weights.
It should be noted that the division of the modules of the above system is only a logical division, and all or part of the actual implementation may be integrated into one physical entity or may be physically separated. And the modules can be realized in a form that all software is called by the processing element, or in a form that all the modules are realized in a form that all the modules are called by the processing element, or in a form that part of the modules are called by the hardware. For example: the x module can be a separately established processing element, and can also be integrated in a certain chip of the system. In addition, the x-module may be stored in the memory of the system in the form of program codes, and may be called by one of the processing elements of the system to execute the functions of the x-module. The other modules are implemented similarly. All or part of the modules can be integrated together or can be independently realized. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in the form of software. These above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), one or more microprocessors (DSPs), one or more Field Programmable Gate Arrays (FPGAs), and the like. When a module is implemented in the form of a Processing element scheduler code, the Processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. These modules may be integrated together and implemented in the form of a System-on-a-chip (SOC).
EXAMPLE III
The present embodiment provides an electronic device, including: a processor, memory, transceiver, communication interface, or/and system bus; the memory is used for storing computer programs and the communication interface is used for communicating with other devices, and the processor and the transceiver are used for running the computer programs so that the electronic equipment can execute all steps of the recommended method of the high school library.
The above-mentioned system bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The system bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus. The communication interface is used for realizing communication between the database access device and other equipment (such as a client, a read-write library and a read-only library). The Memory may include a Random Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the Integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, or discrete hardware components.
The protection scope of the recommendation method for libraries in colleges and universities according to the present invention is not limited to the execution sequence of the steps listed in this embodiment, and all the solutions implemented by adding, subtracting, and replacing the steps in the prior art according to the principles of the present invention are included in the protection scope of the present invention.
The invention also provides a recommendation system for libraries in colleges and universities, which can realize the recommendation method for libraries in colleges and universities, but the implementation device for the recommendation method for libraries in colleges and universities, which is disclosed by the invention, comprises but is not limited to the structure of the recommendation system for libraries in colleges and universities, and all structural modifications and substitutions in the prior art, which are made according to the principle of the invention, are included in the protection scope of the invention.
In summary, the recommendation method, system, readable storage medium and electronic device for libraries in colleges and universities of the present invention provide accurate book recommendation by combining the relevance between student subjects and book types and individual diversity, which not only can make full use of library resources, but also is helpful for improving the learning score of students. The invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (8)
1. A recommendation method for libraries in colleges and universities is characterized by comprising the following steps:
capturing individual preferences of students on books of different classes, and capturing professional preferences of students on books of different classes; receiving discrete attribute data related to a student borrowing a book; mapping the discrete attribute data into dense attribute data related to the book borrowed by the student; establishing an interaction model of each pair of attribute features in the dense attribute data; for a given multi-domain feature x ∈ R n The interactive model modeling mode is as follows:
wherein, ω is 0 Representing a global deviation, ω i Represents the weight value of the ith variable,<v i ,v j >representing the inner product of two vectors, v i ∈R d Is an embedded vector of the ith feature, d denotes the size of the vector, x i Ith characteristic vector value, x j Is the jth feature vector value;
selecting an attribute feature related to the college from a plurality of attribute features; forming a feature interaction graph of attribute features related to the college; convolving the feature interaction graph to form a new multi-channel feature graph; the new multi-channel feature map is used for representing the correlation between every two attribute features; mapping the correlation between every two attribute features into professional preference which represents students are specialized in a specific school period;
and establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, so as to recommend the books matched with the preferences of the students.
2. The recommendation method for libraries at colleges and universities of claim 1, wherein the step of capturing personalized preferences of individual students for different classes of books further comprises:
receiving historical borrowing records of students;
establishing an interactive model of attribute characteristics related to book borrowing of students according to historical book borrowing records of the students;
and establishing the weighted preference of all the borrowing categories of the students according to the historical borrowing records of the students.
3. The recommendation method for libraries at colleges and universities of claim 2, wherein the step of capturing personalized preferences of individual students for different classes of books further comprises:
the weighted preference of all the borrowing categories of the students is interacted with the attribute characteristics to form an interaction result;
and combining the interaction result with the interaction model of each pair of attribute characteristics to obtain the personalized preference of the individual student to the books of different classes.
4. The recommendation method for libraries in colleges and universities of claim 1, wherein the step of convolving the feature interaction graph to form a new multi-channel feature graph comprises:
inputting the feature interactive map of the attribute features related to the college into a convolution layer of a convolution neural network, using a two-dimensional convolution kernel, adding a bias, and performing nonlinear transformation by using an activation function to form a new multi-channel feature map.
5. The recommendation method for library of colleges and universities according to claim 1,
the preference model of the student = (1-lambda). Times the individual student's personalized preference for different classes of books + lambda. Times the professional preference of the student for a particular period; λ is a hyper-parameter controlling two interest weights.
6. A recommendation system for a library of colleges and universities, comprising:
the first processing module is used for capturing the personalized preferences of students on books of different categories; receiving discrete attribute data related to a student borrowing a book; mapping the discrete attribute data into dense attribute data related to the book borrowed by the student; establishing an interaction model of each pair of attribute features in the dense attribute data; for a given multi-domain feature x ∈ R n The interactive model modeling mode is as follows:
wherein, ω is 0 Representing a global deviation, ω i Represents the weight value of the ith variable,<v i ,v j >representing the inner product of two vectors, v i ∈R d Is an embedded vector of the ith feature, d denotes the size of the vector, x i Ith characteristic vector value, x j Is the jth eigenvector value;
the second processing module is used for capturing professional preferences of students on books of different categories; selecting an attribute feature related to the college from a plurality of attribute features; forming a feature interaction graph of the attribute features related to the college; convolving the feature interaction graph to form a new multi-channel feature graph; the new multi-channel feature map is used for representing the correlation between every two attribute features; mapping the correlation between every two attribute features into professional preference which represents students are specialized in a specific school period;
and the comprehensive processing module is used for establishing a preference model of the students according to the individual preferences of the individual students on the books of different categories and the professional preferences of the students on the books of different categories, so as to recommend the books matched with the preferences of the students.
7. A readable storage medium on which a computer program is stored, the program, when executed by a processor, implementing the recommendation method for a library of colleges and universities of any one of claims 1 to 5.
8. An electronic device, comprising: a processor and a memory;
the memory is used for storing a computer program, and the processor is used for executing the computer program stored by the memory to enable the electronic equipment to execute the recommendation method of the library of colleges and universities as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011096851.2A CN112163161B (en) | 2020-10-14 | 2020-10-14 | Recommendation method and system for college library, readable storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011096851.2A CN112163161B (en) | 2020-10-14 | 2020-10-14 | Recommendation method and system for college library, readable storage medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112163161A CN112163161A (en) | 2021-01-01 |
| CN112163161B true CN112163161B (en) | 2022-11-29 |
Family
ID=73868222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011096851.2A Active CN112163161B (en) | 2020-10-14 | 2020-10-14 | Recommendation method and system for college library, readable storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112163161B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113590945B (en) * | 2021-07-26 | 2023-07-28 | 西安工程大学 | Book recommendation method and device based on user borrowing behavior-interest prediction |
| CN113590970B (en) * | 2021-08-11 | 2022-03-22 | 东北林业大学 | Personalized digital book recommendation system and method based on reader preference, computer and storage medium |
| CN116991979B (en) * | 2023-09-27 | 2023-12-01 | 中国科学院文献情报中心 | Matching method and device based on explicit semantic content |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021153A (en) * | 2014-05-19 | 2014-09-03 | 江苏金智教育信息技术有限公司 | Personalized recommendation method for campus books |
| CN106202184A (en) * | 2016-06-27 | 2016-12-07 | 华中科技大学 | A kind of books personalized recommendation method towards libraries of the universities and system |
| CN106802915A (en) * | 2016-12-09 | 2017-06-06 | 宁波大学 | A kind of academic resources based on user behavior recommend method |
| CN109472286A (en) * | 2018-09-30 | 2019-03-15 | 浙江工业大学 | Books in University Library recommended method based on interest-degree model Yu the type factor |
| CN110210927A (en) * | 2019-05-07 | 2019-09-06 | 长沙曙通信息科技有限公司 | A kind of IT books recommender system design based on collaborative filtering |
| CN111612993A (en) * | 2020-05-29 | 2020-09-01 | 安徽文香信息技术有限公司 | Wisdom campus library management system based on internet |
-
2020
- 2020-10-14 CN CN202011096851.2A patent/CN112163161B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021153A (en) * | 2014-05-19 | 2014-09-03 | 江苏金智教育信息技术有限公司 | Personalized recommendation method for campus books |
| CN106202184A (en) * | 2016-06-27 | 2016-12-07 | 华中科技大学 | A kind of books personalized recommendation method towards libraries of the universities and system |
| CN106802915A (en) * | 2016-12-09 | 2017-06-06 | 宁波大学 | A kind of academic resources based on user behavior recommend method |
| CN109472286A (en) * | 2018-09-30 | 2019-03-15 | 浙江工业大学 | Books in University Library recommended method based on interest-degree model Yu the type factor |
| CN110210927A (en) * | 2019-05-07 | 2019-09-06 | 长沙曙通信息科技有限公司 | A kind of IT books recommender system design based on collaborative filtering |
| CN111612993A (en) * | 2020-05-29 | 2020-09-01 | 安徽文香信息技术有限公司 | Wisdom campus library management system based on internet |
Non-Patent Citations (1)
| Title |
|---|
| "基于网络特征学习的个性化推荐系统";王鸿伟;《中国优秀博硕士学位论文全文数据库(博士)信息科技辑》;20200115(第01期);第68-72页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112163161A (en) | 2021-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Ekt: Exercise-aware knowledge tracing for student performance prediction | |
| Pardos et al. | Connectionist recommendation in the wild: on the utility and scrutability of neural networks for personalized course guidance | |
| CN112163161B (en) | Recommendation method and system for college library, readable storage medium and electronic equipment | |
| Liu et al. | Multi-perspective social recommendation method with graph representation learning | |
| Hersche et al. | A neuro-vector-symbolic architecture for solving Raven’s progressive matrices | |
| Tamayo | SmartParticipation | |
| Hong et al. | Selective residual learning for visual question answering | |
| Kardan et al. | A hybrid approach for thread recommendation in MOOC forums | |
| Dien et al. | An approach for learning resource recommendation using deep matrix factorization | |
| Lang et al. | Movie recommendation system for educational purposes based on field-aware factorization machine | |
| Yang et al. | Corporate relative valuation using heterogeneous multi-modal graph neural network | |
| CN112131261A (en) | Community query method and device based on community network and computer equipment | |
| Xu et al. | Educational data mining: dropout prediction in XuetangX MOOCs | |
| Liu et al. | Deep learning and collaborative filtering-based methods for students’ performance prediction and course recommendation | |
| Zhang et al. | Deep pairwise hashing for cold-start recommendation | |
| Du et al. | POLAR++: active one-shot personalized article recommendation | |
| Feng et al. | Ontology semantic integration based on convolutional neural network | |
| Bergner et al. | Multidimensional item response theory in the style of collaborative filtering | |
| Patel et al. | Recommendation Systems: Types, Applications, and Challenges | |
| Sahid et al. | Categorizing attributes in identifying learning style using rough set theory | |
| Nayani et al. | Combination of Deep Learning Models for Student’s Performance Prediction with a Development of Entropy Weighted Rough Set Feature Mining | |
| Nguyen et al. | A knowledge graph embedding based approach for learning path recommendation for career goals | |
| George et al. | Hy-MOM: Hybrid recommender system framework using memory-based and model-based collaborative filtering framework | |
| Niu et al. | A coupled user clustering algorithm based on mixed data for Web-based learning systems | |
| Slaninová et al. | Analysis of social networks extracted from log files |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |