A higher-performance big data-based movie recommendation system

Lin Zhu; Li Zhuang

doi:10.1515/nleng-2025-0092

Enjoy 40% off

academic books on De Gruyter Brill *

Article Open Access

A higher-performance big data-based movie recommendation system

Lin Zhu and Li Zhuang

Published/Copyright: June 4, 2025

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information

From the journal Nonlinear Engineering Volume 14 Issue 1

Abstract

This paper proposes a Movie Recommendation System (MRS) based on big data. MRS adopts B/S architecture and uses front-end and back-end separation modes. The front-end uses the Vue progressive framework, and the back-end uses the SpringBoot back-end development framework combined with MySQL database development. First, the information about the film is crawled from the network by Python crawler technology. Second, the actual needs of the relevant movie management and recommendation system are provided, using SpringBoot, Vue, MySQL, and other technologies equipped with popular movie management, review management, user management, recharge management, and other functions. In addition, this MRS provides a personalized service, recommending films of interest to users based on their preferences. The relevant tests show that MRS has reliable technical support in the later promotion and operation, providing users with a good experience and also laying a solid foundation for system upgrading.

Keywords: big data; recommendation; movie recommendation system; SpringBoot framework

1 Introduction

With the popularity of the Internet and the advent of the information age [1,2,3,4,5], it has become increasingly challenging for users to find content that matches their personal preferences [6,7,8] in the vast ocean of information. In this context, the recommendation system has become a key tool for enhancing user experience and promoting information consumption [9,10,11,12]. These systems analyse a user’s historical behaviour, preferences, and interests and use algorithmic models to predict what items the user is likely to like in order to provide personalised recommendations. In areas such as e-commerce, social media, and online video, recommendation systems have become a key competitive advantage in attracting and retaining users. Movie Recommendation System (MRS) represents an important application area within the field of recommendation systems [13,14,15,16]. Films demonstrate diversity and subjectivity, and user preferences vary widely. Therefore, it becomes crucial to design an efficient and accurate movie management and recommendation system. In order to meet the user’s needs and get the film that suits their preferences, this study designs an MRS based on big data to provide personalized services to users more accurately.

In-depth analysis and mining using film-related data can reveal the potential needs and preferences of users and provide strong support for movie management and recommendation systems. Traditional movie recommendation is often based on popular tastes and popular films, which cannot meet the personalized needs of users [17,18,19]. The movie management and recommendation system based on big data analysis can provide personalized movie recommendations for each user with the help of recommendation algorithms, etc. to enhance user experience and satisfaction.

The existing MRS focuses on user satisfaction, viewing volume, movie exposure, and user stickiness while providing data support for producers. However, for those new users or movies, it may be difficult to be recommended due to insufficient recommendation data. In addition, this MRS limits the discovery of content, raises privacy concerns, exhibits algorithmic bias, and may harm market diversity due to excessive dependence.

The significance of this study is mainly reflected in two aspects: improving the efficiency of movie management and enhancing the user experience. First, the movie management and recommendation system based on big data analysis can help administrators better manage movie inventory, etc., and improve management efficiency and accuracy. Second, by analysing users’ interests and preferences, the system can tailor movie recommendations for each user, providing a personalized viewing experience and enhancing user retention and loyalty to the platform.

In addition, the film management and recommendation system based on big data analysis is also of great significance to the development of the film industry. By analysing users’ preferences and movie-watching behaviours, the system can provide targeted market analysis and decision-making support for film practitioners and promote the innovation and development of the film industry. At the same time, the research results of this study have the reference value of academic research, which can provide reference for related fields and safeguard the potential application of film management and recommendation systems in other fields.

2 Theory and technology

This section mainly introduces the details of the front-end technology, back-end technology, database design, and content-based recommendation algorithm to support the proposed MRS.

2.1 Front-end technology

2.1.1 The Vue.js (Vue) framework

Vue.js [20,21,22] is an incremental framework that focuses on the view layer and works with useful component libraries like ElementUI for interaction logic and page design. Components are an important feature of Vue.js. The Vue component system implements extended Hyper Text Markup Language (HTML) elements that encapsulate the available code. Vue.js mainly uses the MVVM pattern, encourages the use of HTML template rendering, and features two-way data binding. In addition, Vue.js has a rich plugin ecosystem that makes it easy to integrate third-party plugins and libraries with a simple and flexible API. This gives it the ability to create data-driven user-interface components.

In this system, Vue.js technology is used to divide the page into components such as movie search and movie enjoyment. Each component is independent, effectively reducing code redundancy and duplication. In the movie viewer component, Vue.js can retrieve movie data using asynchronous loading for faster page loads and a smoother user experience.

Vue.js also has its limitations. It is a complex framework that new developers may need time to learn, which increases the learning cost. Modularization helps with code reuse and maintenance, but improper component partitioning and communication design may lead to code confusion and increased system complexity. Asynchronous loading of Vue.js can improve speed, but it may face performance challenges when processing large amounts of data or complex logic, requiring optimization. These issues have been fully considered in this work.

2.1.2 HTML technology

HTML [23,24,25,26] is one of the technologies required for front-end development and is one of the most widely used markup languages in web content creation and authoring technologies today. HTML is a structured markup language for web page content, with tags and attributes making up the entire page. In this system, HTML is used to create a visual interface, including a user input interface and a movie information display interface, which is easy for users to visualise and use. In addition, the clean web design improves the overall system, and the minimalist web design improves the user experience.

It should be noted that HTML has limited interactive and dynamic data processing capabilities, so it is usually necessary to combine Cascading Style Sheet (CSS) and JavaScript to create fully functional web pages.

2.1.3 CSS technology

Use CSS technology [27,28] to design the layout, colour, font, size, and other appearance styles of web pages. The technology provides a number of different layout styles from which designers can choose their preferred settings, including flow layouts, floating layouts, grid layouts, and more. Fonts and text styles can also be adjusted to your liking in terms of colour, size, and alignment to beautify the layout. Background images and borders can also be set and adjusted. In addition, CSS can complete the transition and transformation between different state elements, including rotation, scaling, and so on, so as to enhance the user experience.

Although CSS technology is powerful, it takes time for beginners to master it. Different browsers have differences in parsing, resulting in inconsistent display and requiring additional debugging. In addition, excessive use may slow down webpage speed and affect the user experience. Therefore, designers should weigh the pros and cons and use CSS reasonably.

2.2 Back-end technology

2.2.1 SpringBoot framework

SpringBoot framework [29,30,31] is an open-source application framework on the Java platform that provides containers with control inversion features. Although the framework itself has no restrictions on the programming model, its frequent use in Java applications has made it so popular that it has since been made to complement, if not replace, the Enterprise Java Beans model. The framework has very rich third-party libraries, such as Redis connection pool and database connection library. The back-end of this system uses the idea’s development environment and uses Maven to manage the project, thus making the configuration and development of the system easier.

2.2.2 MyBatis architecture

MyBatis [32,33,34] is a persistence layer framework. It supports custom SQL, stored procedures and advanced mapping, avoiding almost all the JDBC code and manually set parameters and get the result set. MyBatis can be configured and mapped using simple XML or annotations to native information.

The MyBatis architecture is divided into three layers: the API interface layer, which allows developers to operate the database; the data processing layer handles SQL lookup, parsing, execution, and result mapping; and the basic support layer manages connections, transactions, configurations, and caching. The main drawbacks are SQL efficiency issues in the data processing layer and cache configuration risks in the underlying support layer.

2.3 Database design

This system uses the commonly used MySQL 5.7 [35,36] for database design. This database will transmit passwords in an encrypted form to ensure data security and meet development requirements.

This system will use MySQL to store the basic information of the user, including user name, gender, login password, etc. It will also be used to store the basic information of the film including the name of the film, the poster, the genre, the name of the actor and the director, the duration, etc. This information is crawled out from the Douban website using crawler technology. At the same time, user clicks, browsing, and other behavioural data generated on the front page of the system are also stored in the database. By analysing this data, the user’s preferences and biases can be further understood.

2.4 Content-based recommendation algorithm

This system can determine the user’s preferred film type according to the number of clicks on popular films, thus displaying the recommended module on the home page. As an example, a user likes a certain type of film, the user logs into the system, and based on the number of clicks viewed under that type of film in the popular films module, the number of clicks on that type of film by that user is aggregated and sorted by bubbling. Further, the cosine similarity value is calculated, and a similarity comparison is performed. Finally, based on the bubble sorting, the same type of recommendation is performed to determine the user’s favourite movies of the genre so that the movies under the genre are continuously recommended to the user.

The principle of this recommendation algorithm [37,38] is shown in Figure 1.

Figure 1

Schematic diagram of the adopted recommendation algorithm.

For the cosine distance algorithm, the closer the cosine value is to 1, the closer the angle is to 0 degrees, i.e. the more similar the two vectors are, which is called cosine similarity. The range of similarity is [−1, 1]. If the cosine value is close to −1 then it is not similar. If the angle between the a and b vectors is small, then the a and b vectors have a high degree of similarity. In extreme cases, the a and b vectors will coincide exactly. The cosine similarity is calculated using Eq. (1):

(1) cos ( u , v ) = r u ⋅ r v | | r u | | ⋅ | | r v | | = ∑ i = 1 n r u , i ⋅ r v , i ∑ i = 1 n ( r u , i ) 2 ⋅ ∑ i = 1 n ( r v , i ) 2 ,

where r _u denotes the preference profile of a particular user for a film, r _v denotes the preference profile of a particular candidate film, r u ⋅ r v denotes film u -to-film v multiplication, and | | r u | | ⋅ | | r v | | denotes taking the square and then the square root of the preference feature for the film u multiplied by taking the square and then the square root of the preference feature for the film v .

The recommendation process can be categorised into the following four scenarios.

If you don’t get the logged-in user, it will return the default list data directly.
If you get a logged-in user, first look at the user’s most frequent clicks to see if there is any data. If there is no data, it means that the current user has no click data for the time being. At this point, return to the default list of data, that is, add time before the first 12 data as SELECT * FROM popular_movies ORDER BY create_time DESC LIMIT 0,12.
Calculate the similarity between film types. If there is click data from the user, it is sorted after taking out the data to get the maximum value of clicks (hits count). Query the data under that category based on the click category with the most user clicks calculateCosineSimilarity (typeName, String. value of (typeCos.get (“movie genre”))). If there is insufficient data, the cosine similarity dot Product/(magnitude1 * magnitude2) is calculated for the other type and the user’s current most clicked. At this point, the data 12 - resultList.size()) is obtained sequentially downwards, and according to the matching degree of the type and the bubble sorting method, the data is obtained from the highest to the bottom so as to recommend the film to the user.
If there are more than or equal to 12 pieces of data, the data is bubble sorted, and the 12 pieces of data with the most hits are returned.

The flowchart of the recommendation process is shown in Figure 2.

Figure 2

Flowchart of the recommendation process.

3 System development

3.1 Architectural design

The system uses a B/S structure (Browser/Server) [39,40,41]. The architecture decouples web services and application services. First, an Http request is made from the client to the web server. The web server handles the Http request and calls the RESTFUL interface exposed by the application server. If it is necessary to interact with the database, the application server will interact with the database and return the relevant Json data to the web server. At this point, the web server renders the template and data combination into the form of HTML and returns it to the client. The B/S architecture workflow of this system is shown in Figure 3.

Figure 3

B/S architecture workflow diagram.

3.2 Functional module design

Ordinary users can enter the user module for personal information management, film browsing, film search, film collection, film comments, etc. after logging in. New users can register to get an account and log in to the system to view and edit their personal information, browse movie information, search for movies, etc. They can also like and favourite the movies they like and comment on the movies they have watched. The system will recommend popular films on the homepage according to individual clicking preferences. Administrators can access the administrator module after logging in backstage to manage users and administrators, manage users’ recharge balance and recharge records, and check the paid reviews of films. In addition, they can also manage popular films, resources and system information.

3.2.1 Popular movie module

In the popular movie module, users can find their favourite popular movies by searching for the genre, movie title, or actor name. The interaction between the platform and the user is improved by pushing films with similar types of clicks to the user’s homepage. Meanwhile, users can like, favourite, and comment on their favourite popular movies. Users can reply to each other, share their thoughts and opinions, and enhance the connection between users. In addition, users can watch the corresponding film by deduction, the cost of which is deducted from the user’s top-up amount, enhancing the user’s experience.

On the left side of the popular film details interface, information such as name, genre, director, actors, release date, and duration are provided, while on the right side of the interface are the posters of the corresponding films. At the bottom of the poster, there are watch, like, and favourite buttons, which allow users to like or favourite their favourite movies to express their love for them. At the same time, you can pay for the films you are interested in by clicking the watch button. In the area below the green dotted line are the comment display area and the comments area. Users can post comments in the comments section, while the display area presents the comments posted by users. In this way, users can select a user comment to reply to and share their views with other users. This design makes it easy for users to learn about others’ views and opinions on films and to participate in interactive activities such as liking, favouriting, and commenting on popular films.

3.2.2 Film information module design

In the film information module, users can read and learn about the latest film information, including the content of newly released films, film news information and so on. Users can like and favourite the content they like, and at the same time, they can comment on the corresponding film information to express their views.

The left side of the film information interface displays the main information picture and title, and the white list on the right side mainly reflects the content of the corresponding information. Users can learn more about the content by clicking on it. Users can use the search box for fuzzy queries to quickly find the film information they are interested in or click on the filter or sort button to find and browse the information in a purposeful way. This interface helps users to understand the film information and makes it easy for them to choose the film information they like to watch.

3.2.3 User management module

In the background, managers can query the user information to edit the front-end user’s relevant information data. Users can also be cancelled. Adding and deleting administrators’ editing administrator information is also permitted, which makes the management of system user personnel more tidy and effectively strengthens personnel management.

In the film management and recommendation system, managers can control the registered ordinary users and administrators through the System Users menu. Users can view the list of administrators and users, including nicknames, usernames, and registration dates, and click the green details button in the details column to further view user and administrator profiles. Through the system user function, managers can effectively control the registered ordinary users and administrators to ensure the security, stability, and user experience of the system.

3.2.4 Balance information management module

The management of user balance information facilitates the administrator to quickly understand the remaining amount of recharge money of each user. The administrator can check the user’s balance information according to the user’s name, and the incorrect chargeback balance information can be modified and edited. This enhances the thoughtfulness of the system management and gives the user a good user-centred experience.

The user list shows the user account, user name, user phone number, top-up time and amount and the top-up details filled by the user.

3.2.5 Top-up records management module

The main purpose of recharge record management is to count the records of recharged users and present the relevant data in the form of statistical charts on the home page of the back-end. If more users are willing to recharge to watch popular movies, then it reflects that the popular movie content provided by the system front-end is satisfactory and popular to users. The query by user name and time of top-up facilitates the administrator to obtain information about the top-up records of different users in different time periods. In addition, the administrator can change the wrong recharge amount from the background.

3.2.6 Pay-per-view management module

The main purpose of pay-per-view management is to count the amount of money spent by the front-end users on film viewing and present the relevant statistics in the form of statistical charts on the back-end homepage. This facilitates the administrator to intuitively understand the income of the film, which is conducive to the subsequent management and updating of popular films and improves the user’s satisfaction with the use of the site.

3.2.7 Review statistics management module

The main purpose of point statistics management is to count the number of likes and comments users have made on different popular movies in the front-end. The list of this module consists of the name of the film, its genre, the time of the statistics, and the number of likes and comments corresponding to the popular films. The related data is displayed on the background homepage in the form of statistical charts. In the backstage, first, we integrate and get the data of likes and comments of each popular film, and then download and import the document to save as a file in the interface of the review statistics list. Add the film data as per the format on the document, and finally, import the data by clicking on the import button to generate the latest likes and comments information related to that film. This approach makes it easier for administrators to have a more intuitive understanding of the number of user likes and comments on different popular movies, thus understanding user preferences, which helps administrators manage the follow-up of popular movies.

3.2.8 Popular movie management module

The hot movies management module is an important part of the movie management and recommendation system, which is used for administrators to manage and display hot movies. This module includes the functions of popular movie searching, popular movie adding, popular movie editing, popular movie deleting, popular movie statistics, and popular movie recommending. Administrators can view the list of all popular movies and search and filter movies on demand. Administrators can add new information about popular movies, edit existing information, and delete movies that do not meet the criteria. This module facilitates the administrator to maintain and update the list of popular movies, providing users with an accurate and attractive showcase page of popular movies.

3.2.9 Resource management module

In order to ensure the quality of film information content, administrators can edit, add, or delete film information in the resource management module to update film-related information and specific message article content in real time to enhance user satisfaction. At the same time, a comment management section has been created to effectively control undesirable information. Administrators can delete undesirable comment content to create a friendly system comment environment. At the same time, this ensures that the system can effectively store and manage movie information, provide accurate and personalised recommendation services and facilitate user and administrator interaction and management. In addition, administrators can upload, edit and delete resources through this interface, in which the MultipartFile method [42,43] is used for image upload.

3.2.10 System management module

The system management module is mainly used for the management of rotating pictures on the front home page. In this module, administrators can add, delete and modify the rotating image and replace the film promotional image that fits the real-time information. In this way, users can understand the current hot film news more intuitively on the home page, providing users with an accurate and attractive home page rotator display page.

Figure 4

Example of charging record statistics chart.

Figure 5

Example of a statistical chart of popular films.

Figure 6

Example of viewing payment statistics chart.

Figure 7

Example of a statistical chart of points of view.

3.2.11 Back-end home page interface

The homepage interface of the background is mainly presented by statistical charts, in which the recharge record statistical chart is a line chart that can reflect the user’s recharge situation for different films. The popular films statistics chart is a pie chart that can more intuitively see the popularity of popular films. The comment statistics chart is a bar chart that can directly show the users’ likes and comments on the films. This data presentation facilitates the administrator’s film management and makes it clear to the administrator the popularity of different types of films. Call the component by importing echarts from @/assets/js/echarts.min.js for statistical chart drawing.

The top-up record statistics graph (Figure 4) counts the top-up amounts of different users at different times. The horizontal axis of the line graph represents the time of top-up, and the vertical axis represents the amount of top-up. Different colour circles at the bottom represent different users; for example, dark blue represents 123 users, and the managers will click the mouse on the corresponding folding line will pop up the number of different users at the same time, which is helpful for the managers to intuitively understand the user’s recharge situation.
The popular films statistical chart (Figure 5) counts the percentage of different film genres in the pie chart. The left side shows the different colour blocks representing the genres of films, e.g. dark blue blocks represent biographical films. On the right side are pie chart statistics corresponding to the percentage share of films of different genres, and each block in the chart will indicate the corresponding genre. When the administrator clicks on the different coloured pie charts, the specific share data for that genre of film is displayed.
The pay-per-view statistics graph (Figure 6) uses dark blue bars to count the revenues from being topped up for different films. The vertical axis of the chart represents the total number of top-ups for the corresponding film, and the horizontal axis represents the name of the film. Administrators can visually see the total amount of top-ups for different films.
The points statistics graph (Figure 7) also shows the number of likes and comments for different films using bar graphs. The blue coloured blocks at the bottom of the stats graph represent the number of likes, and the green coloured blocks represent the number of reviews. The vertical axis represents the number of reviews and comments, and the horizontal axis represents the name of each film. Administrators can click on the bars of different colours to display the name of the movie, number of likes, and comments.

4 Testing and analysis

4.1 System test environment

The operating system is Windows 11, with a 2.0 GHz CPU, 8GB of RAM, and a 100 MB hard drive. An image-grade graphics card is used, and there are no GPU requirements.

4.2 System testing process

The testing process is as follows.

Determine testing objectives: Clarify the performance indicators that MRS needs to achieve, such as accuracy, recall, coverage, etc.
Prepare test data: Collect or generate movie rating data, user behavior data, etc., for testing purposes.
Design test cases: Design test cases based on different scenarios, including user history behavior, ratings, preferences, etc.
Implement testing environment: Build a testing environment similar to the production environment to ensure accurate execution of test data and test cases.
Execute test cases: Run test cases, collect recommended results, and compare them with expected results.
Performance evaluation: Use statistical analysis methods to evaluate the performance indicators of MRS, such as accuracy, recall, etc.
User experience testing: Invite real users to participate in the test and collect their satisfaction feedback on the recommendation results.
Defect repair and optimization: Based on test results and user feedback, fix identified issues and optimize recommendation algorithms.
Regression testing: After defect repair and optimization, re-execute test cases to ensure that improvements do not introduce new issues.
Test report: Write a detailed test report summarizing the testing process, results, and performance evaluation of MRS.

4.3 System test cases

The system test includes a user login function test, movie information display view function test, popular movie add, popular movie search, password change function test, and popular movie push test, as shown in Tables 1–6, respectively.

Table 1

User login function test table

Use case name	User login system
Goal	Test whether users can log in with the correct username and password
Prerequisite	Not logged in
Testing process	Enter the login screen Enter the correct user and password
Expected result	If your username and password are correct, you will be redirected to the login screen; otherwise, an error message will be displayed and you will be prompted to re-enter it
Actual result	Actual results are in line with expected results

Table 2

Movie information viewing function test table

Use case name	Film news view
Goal	Test movie information viewing function
Prerequisite	User login
Testing process	Click on the film information list
Expected result	You can view all the film information
Actual result	Actual results are in line with expected results

Table 3

Hot movie add test table

Use case name	Top movies added
Goal	Test hot movie adding function
Prerequisite	Administrator users logging in normally
Testing process	The administrator clicks on popular movies and then clicks on add after and fills in the information Click to submit
Expected result	After submitting, new film information will be displayed on the home page
Actual result	Actual results are in line with expected results

Table 4

Popular movie search test table

Use case name	Popular movie search
Goal	Test the search function for popular films
Prerequisite	None
Testing process	Fill in the search box with the search keywords Click on the search button
Expected result	The page displays popular films that contain the search keywords
Actual result	Actual results are in line with expected results

Table 5

Password change function test table

Use case name	Password change
Goal	Test administrator password change function
Prerequisite	Administrator users logging in normally
Testing process	Change the administrator password and complete the form Click to submit
Expected result	Logging in is possible with a new password
Actual result	Actual results are in line with expected results

Table 6

Popular movie recommendation test table

Use case name	Hot movie recommendation
Goal	Whether the platform pushes similar film genres based on user clicks
Prerequisite	Users logging in normally
Testing process	Users search and tap on their favourite movies on the popular movie screen Click to submit
Expected result	The home page pushes similar film genres to users, and the film with the highest number of user views and likes is in the first place
Actual result	Actual results are in line with expected results

By writing test cases for the film management and recommendation system, it completes the user login function test, film information display, and view function test, popular film add, popular film search, and password change function test, which provides strong technical support for the system’s later promotion and operation. These tests ensure the proper functioning and stability of the system on the five key modules. User login functionality tests verify that users are able to successfully register, log in, and logout to ensure security. The movie information display viewing functionality test checks the accuracy of movie information and the consistency of the page display. The popular movies add test ensures that administrators can successfully add and manage popular movies to provide engaging content. Popular movie search tests verify that users can quickly find movies of interest based on keywords or filters. The password change feature tests ensure that users can change their passwords securely. Through these tests, the proposed system is equipped with reliable technical support in the later promotion and operation, providing users with a good experience and laying a solid foundation for the development and upgrading of the system.

5 Summary and outlook

In this study, we adopt the currently popular Spring Boot framework, the Vue front-end framework, and MySQL database technology. By utilizing two powerful programming languages, Java and Python, we design and implement an MRS based on big data analysis. The main goal of this system is to provide a rich source of films for various users, thereby enriching the spiritual and cultural life of community residents.

During the design and development process, we place particular emphasis on the movie management and recommendation functions, considering them as the core features of the system. The movie management function includes operations such as entering, editing, deleting, and querying movie information, making it convenient for administrators to manage the movie database. The movie recommendation function employs big data analysis techniques to intelligently recommend films that users may be interested in based on their viewing history, preferences, and behavior patterns. Through the organic combination of these two core functions, our system not only improves the efficiency of movie management but also greatly enhances the viewing experience for users.

MRS needs to pay attention to data encryption and privacy protection. It is recommended to adopt advanced encryption technology to ensure data security and prevent sensitive information from being illegally accessed. MRS should comply with data privacy policies, clarify data processing rules, and enable users to control personal information. In addition, it is necessary to conduct regular security audits to ensure the implementation of privacy policies and update them to address new threats and regulations.

In terms of the recommendation function, personalized and accurate movie recommendations are provided through a content-based recommendation algorithm, which improves user experience and satisfaction. In order to achieve access to a huge number of films, Python based crawler technology is used to accurately extract different types of films from the web. In this way, users can like, favourite, comment, and watch according to their preferences on the front-end.

In the future, more data sources can be further introduced, new algorithms and models can be explored, multi-dimensional recommendations can be realized, user participation and interaction can be encouraged, and development and deployment can be carried out on multiple platforms. First, combining deep learning techniques, especially convolutional neural networks, and recurrent neural networks, will help extract complex user preferences and movie features, making recommendation results more intelligent. Second, adopting a hybrid recommendation system strategy that combines content-based recommendation with collaborative filtering can better meet the needs of different users and enhance their satisfaction. In addition, by introducing reinforcement learning algorithms, the system can continuously optimize recommendation strategies in real-time feedback to adapt to users’ real-time changes. Finally, by utilizing big data analysis and natural language processing techniques, we can delve deeper into user comments and rating data, further enhancing the accuracy and richness of recommendations. The combination of these algorithms and methods will bring a more intelligent and personalized user experience to our MRS, ensuring its leading position in the fiercely competitive market. The movie management and recommendation system based on big data analysis is able to provide users with personalized and accurate movie recommendations to enhance user experience. By collecting and analysing data such as users’ historical behaviours, preferences, and evaluations, the proposed system is able to understand users’ preferences and make intelligent recommendations based on the recommendation algorithm.

The system helps users discover new films, expand their viewing range, and provide recommended content related to their interests. The system also provides users with detailed film information and ratings to help them make more informed choices. For film managers, the system provides in-depth data insights and market analyses to help them understand user needs and trends and make decisions accordingly. It also provides statistical reports and data visualisation tools to help managers monitor film performance and effectiveness.

In summary, a movie management and recommendation system based on big data analytics can provide users with personalized movie recommendations, help them discover new movies, and enrich the movie-watching experience. For film managers, such a system provides data-driven decision support and market insights. Therefore, this system has important practical significance and wide application prospects.

Acknowledgments

This research was funded by the General Project for Philosophy and Social Science Research in Higher Education Institutions (Grant No. 2024SJSZ0245), and by the Computer Basic Education Teaching Research Project of the National Computer Basic Education Research Association of Higher Education Institutions (Grant No. 2024-AFCEC-277).

Funding information: This research was funded by the Fundamental Research Fund Project of Southeast University (Grant No. 2242022K30056).
Author contributions: Lin Zhu: conceptualized and designed the study. Collected, analyzed, and interpreted the data. Drafted and revised the manuscript. Li Zhuang*: contributed to critical decisions in study design and execution. Provided key technical guidance on data analysis and interpretation. Assisted in writing and revising the manuscript. All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
Conflict of interest: Authors state no conflict of interest.
Data availability statement: All data generated or analysed during this study are included in this published article.

References

[1] Moon WK, Kahlor LA. Fact-checking in the age of AI: Reducing biases with non-human information sources. Technol Soc. 2025;80:102760.10.1016/j.techsoc.2024.102760Search in Google Scholar

[2] Zhang P, Yang P, Kumar N, Hsu CH, Wu S, Zhou F. RRV-BC: Random reputation voting mechanism and blockchain assisted access authentication for industrial internet of things. IEEE Trans Ind Inform. 2024;1:713–22.10.1109/TII.2023.3271127Search in Google Scholar

[3] Prabhu SG, Jayarajan P, Nandakumar E. Protected and energy-efficient e-healthcare system based on internet of healthcare services. J Circuits Syst Comput. 2024;33(04):2450060. 10.1142/S0218126624500609.Search in Google Scholar

[4] Jia X, Parrott A. Flexible fibres take fabrics into the information age. Nature. 2024;626:38–9.10.1038/d41586-024-00076-xSearch in Google Scholar PubMed

[5] Minchev Z. On the growing transformational role of AI technologies for the future cyber diplomacy in the post-information age. Int J Cyber Dipl. 2023;4:29–41.10.54852/ijcd.v4y202303Search in Google Scholar

[6] Fletcher KK. A method for aggregating ranked services for personal preference based selection. Int J Web Serv Res. 2019;16(2):1–23.10.4018/IJWSR.2019040101Search in Google Scholar

[7] Li Y, Tan R, Gong X. How omnichannel integration promotes customer word-of-mouth behaviors: The mediating roles of perceived personal preference fit and perceived social relatedness. Inf Technol People. 2023;36(4):1726–53.10.1108/ITP-06-2021-0440Search in Google Scholar

[8] Zhang H, Tzempelikos A, Liu X, Lee S, Cappelletti F, Gasparella A. The impact of personal preference-based thermal control on energy use and thermal comfort: Field implementation. Energy Build. 2023;284:112848.10.1016/j.enbuild.2023.112848Search in Google Scholar

[9] Kim Y, Shim K. TWILITE: A recommendation system for Twitter using a probabilistic model based on latent Dirichlet allocation. Inf Syst. 2014;42(Jun):59–77.10.1016/j.is.2013.11.003Search in Google Scholar

[10] West JD, Wesley-Smith I, Bergstrom CT. A recommendation system based on hierarchical clustering of an article-level citation network. IEEE Trans Big Data. 2016;2:113–23.10.1109/TBDATA.2016.2541167Search in Google Scholar

[11] Essay P, Rajasekharan A. Robust diagnosis recommendation system for Primary Care Telemedicine using long short-term memory multi-class sequence classification. Heliyon. 2024;10(6):26770.10.1016/j.heliyon.2024.e26770Search in Google Scholar PubMed PubMed Central

[12] Sivanandam C, Perumal VM, Mohan J. A novel light GBM-optimized long short-term memory for enhancing quality and security in web service recommendation system. J Supercomput. 2024;80(2):2428–60.10.1007/s11227-023-05552-1Search in Google Scholar

[13] Darshini M, Raina A, Lokesh RM, Durrani MNK, Sreenivas TH. Movie recommendation system. IJEAST. 2019;(11).Search in Google Scholar

[14] Marappan R. Movie recommendation system using an item-based collaborative filtering. Int J Math Eng Biol Appl Comput. 2022;1:42–3.10.31586/ijmebac.2022.340Search in Google Scholar

[15] Darban ZZ, Valipour MH. GHRS: Graph-based hybrid recommendation system with application to movie recommendation. Expert Syst Appl. 2022 Aug;200:116850.10.1016/j.eswa.2022.116850Search in Google Scholar

[16] Kunde O, Gaikwad O, Kelgandre P, Damodhar R, Swami MM. The movie recommendation system using content based filtering with TF-IDF-vectorization and Levenshtein distance. Int J Adv Res Sci Commun Technol. 2022;2(2):257–63.10.48175/IJARSCT-3648Search in Google Scholar

[17] Szomszor M, Cattuto C, Alani H, O'Hara K, Baldassarri A, Loreto V, et al. Folksonomies, the semantic web, and movie recommendation. In: Proc. of the 4th European Semantic Web Conference, Bridging the Gap between Semantic Web and Web 2.0. 2007.Search in Google Scholar

[18] Liu N, He L, Zhao M. Social temporal collaborative ranking for context aware movie recommendation. ACM Trans Intell Syst Technol. 2013;4:1–26.10.1145/2414425.2414440Search in Google Scholar

[19] Benini S, Canini L, Leonardi R. A connotative space for supporting movie affective recommendation. IEEE Trans Multimed. 2011;13(6):1356–70.10.1109/TMM.2011.2163058Search in Google Scholar

[20] Fang S. Research based on Vue.js front-end framework technology. Comput Knowl Technol. 2021;17(19):3.Search in Google Scholar

[21] Baida R, Andriienko M, Plechawska-Wójcik M. Performance analysis of frameworks Angular and Vue.js. J Comput Sci Inst. 2020;14:59–64.10.35784/jcsi.1577Search in Google Scholar

[22] Yin G, Li C, Zeng Z, Zhao P, Lei P, Chang L. Dam safety monitoring information management system development based on Vue.js + Django. Comput Eng Softw. 2024;45:45.Search in Google Scholar

[23] Berners-Lee T, Connolly D. Hyper text markup language. Reston, Virginia, USA: Internet Society (ISOC); 1995.10.17487/rfc1866Search in Google Scholar

[24] Wu Y, Xiao L, Ding X, Wang B, Zhang J. Real-time data and visualization monitoring of computer numerical control machine tools based on hyper text markup language 5. J Donghua Univ. 2019;36(3):6.Search in Google Scholar

[25] Yang X, Xing H, Ji X, Su X, Pedrycz W. Multitime scale thunderstorm monitoring system with real-time warning and imaging. IEEE Trans Fuzzy Syst. 2024;32(4):1821–35.10.1109/TFUZZ.2023.3336637Search in Google Scholar

[26] Huo J. Widget technology based on the open source web online computer network simulation experiment system. China Comput Commun. 2023;35(11):121–5.Search in Google Scholar

[27] Lie HW, Bos B. Cascading style sheets: Designing for the web [M]. Cambridge, USA: Addison-Wesley Professional; 2005. Search in Google Scholar

[28] Wang D. The application of cascading style sheet in dynamitic HTML. J Anshan Norm Univ. 2006.Search in Google Scholar

[29] Reddy K, Upadhyayula S. Beginning Spring Boot 3: Build dynamic cloud-native Java applications and microservices. Working with JPA: BSpring Data JPA; 2015. p. 84–97.Search in Google Scholar

[30] Wang Y, Shu H. Design and implementation of a university laboratory appointment management system based on SpringBoot + Vue front-end and back-end separation. Mod Comput. 2023;29(1):114–7.Search in Google Scholar

[31] Wei W, Niu S, Zhang B, Schererline R, Damaševičius R. Design and implementation of discipline competition management system based on SpringBoot + Vue technology. Comput Knowl Technol. 2023;19(10):54–8.Search in Google Scholar

[32] Guo J. Design and implementation of quantitative assessment system for students’ moral education based on SpringBoot + Mybatis + Vue. Mod Inf Technol. 2023;7(1):5.Search in Google Scholar

[33] Luo H, He C, Li J, Wang Z, Liu L. Design and implementation of academic achievement repository for universities based on SpringMVC and MyBatis. Mod Inf Technol. 2024;8:008.Search in Google Scholar

[34] Xiao S, Yang K, Wang Y. A study on the optimisation and application of the MyBatis framework. Eng Technol. 2024;(002).Search in Google Scholar

[35] Yang X, Xing H, Ji X, Zhao D, Su X, Pedrycz W. Multifeature fusion-based thunderstorm prediction system with switchable patterns. IEEE Sens J. 2023;23(16):18461–76.10.1109/JSEN.2023.3291397Search in Google Scholar

[36] Wen Y, Hsu C. A performance evaluation of modular functions and state databases for Hyperledger Fabric blockchain systems. J Supercomput. 2023;79:2654–90.10.1007/s11227-022-04762-3Search in Google Scholar

[37] Hu J, Gao Z, Pan W. Multiangle social network recommendation algorithms and similarity network evaluation. J Appl Math. 2013;2013:1–8.10.1155/2013/248084Search in Google Scholar

[38] Dong Z. Resume recommendation based on text similarity. Appl Comput Eng. 2023;6:719–24.10.54254/2755-2721/6/20230937Search in Google Scholar

[39] Li W. Research on the design of smart teaching system based on B/S structure in the context of smart campus. China Comput Commun. 2024;6:036.Search in Google Scholar

[40] Li Y. Design and realisation of remote condition monitoring and intelligent diagnosis system for petrochemical rotating equipment based on B/S model. J Guangdong Univ Petrochem Technol. 2023;6:71–5.Search in Google Scholar

[41] Sheng S, Li J, Wu X, Hou L, Fan L. Design and realization of the graphic system of power system based on browser/server model. In: Proc. Fourth Int Conf Comput Inform Sci. IEEE Computer Society; 2012.10.1109/ICCIS.2012.123Search in Google Scholar

[42] Xue F, Liang F, Xu S, Ren W. Web research and application based on Spring MVC framework. J Hefei Univ Technol. 2012;35(3):4.Search in Google Scholar

[43] Yang J Realization of cloud sharing mode of smart training system based on intelligent recommendation algorithm of legal clauses. Int J Sci Eng Appl. 2023.Search in Google Scholar

Received: 2024-09-27

Revised: 2025-02-13

Accepted: 2025-02-14

Published Online: 2025-06-04

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

https://doi.org/10.1515/nleng-2025-0092

Keywords for this article

big data; recommendation; movie recommendation system; SpringBoot framework

Creative Commons

BY 4.0