English teaching system design based on ID3 algorithm and node optimization

Feature extraction

The digital teaching resources stored in the computer hard disk are collected and processed through the system’s front-end collecting equipment. The resource processing process is as follows: first, delete the processing. Delete duplicate and shortage of valuable resources; avoid occupying storage space; improve the system’s running speed; delete batch based on JSP pages; and second, word processing. English text in the tense verb changes and irregular changes in the plural number of nouns promote the word to appear in a variety of forms, so in the word processing, the same word in different forms will be treated as different words, resulting in complex feature items.

English multimedia teaching resources are characterized by an extensive vocabulary, necessitating feature extraction, i.e., the automated selection of keywords capable of representing content from these resources. This module aims to filter out words with low or no information, simplify the vector space dimension calculation process, avoid overfitting, and simplify calculations while improving classification accuracy (Yang, 2022). There are several techniques for extracting lexical features, among which this article employs the ID3 information gain method to extract lexical features.

Initially, the characteristics of the learners, resource features, and classification attributes are extracted and stored in a two-dimensional array. Subsequently, the discrete data is subjected to preprocessing to achieve normalization. During the process of adaptive resource push, eight attributes warrant consideration: question type (Th), subject matter (Qt), difficulty (Di), reading ability (Ra), cognitive style (Cs), learning goal (Lt), learning situation (Ls), and learning effect (Re). Each attribute possesses a distinct range of values. The conditional entropy of each attribute is computed individually, along with the information entropy of the remaining attributes under the condition of the classification category (Devasenapathy & Duraisamy, 2017).

Taking attribute Th as an example, based on Eq. (1), the information entropy generated by attribute Th is given by: (1)${\displaystyle {\text{Entropy}}_{\mathrm{Th}}\left(X\right)=\sum _{i=1}^k\frac{\left|X\right|}{\left|X\right|}\text{Entropy}\left(X_i\right)=\frac{\left|X_{Sc}\right|}{\left|X\right|}\text{Entropy}\left(X_{Sc}\right)+\frac{\left|X_{Ns}\right|}{\left|X\right|}\text{Entropy}\left(X_{Ns}\right)=\frac{n_c}{n}\left[-\frac{n_m}{n_c}log2\left(\frac{n_m}{n_c}\right)-\frac{n_n}{n_c}log2\left(\frac{n_n}{n_c}\right)\right]+\frac{n_s}{n}\left[-\frac{n_i}{n_s}log2\left(\frac{n_i}{n_s}\right)-\frac{n_j}{n_s}log2\left(\frac{n_j}{n_s}\right)\right].}$

The information gain of Th attribute is shown in Eq. (2): (2)${\displaystyle \text{Gain}\left(\text{Th}\right)=\text{Entropy}\left(X\right)-{\text{Entropy}}_{\mathrm{Th}}\left(X\right).}$

In this way, the value of each node is used to construct the maximum value of the multimedia resource.

Resource sharing process

Through the resource sharing module, the system provides users with remote use of resources on the machine, reducing the waste of resources in the process of sharing. Figure 3 shows the sharing process of the whole English multimedia teaching resource-sharing platform.

In the process of peer-to-peer resource sharing, the peer-to-peer resource sharing platform is used to achieve the goal of the process. There are three nodes in the module. Users can select nodes to perform tasks at any time (Rahman, Newaz & Au, 2020).

First, each node can play two roles at the same time, which are resource users and resource providers. Other remote nodes can run jobs of this node, and this node can also execute jobs of other remote nodes.

Then, when a node is a submission node, it can detect, access special editors, and count reputation. These reports exist in peer-to-peer networks and can be accessed and viewed by all nodes (Sun, Yu & Fan, 2020).

Finally, when a node assumes the role of an execution node, it is equipped with a reliable Java Virtual Machine (JVM) that facilitates the transmission of application progress information to reports.

Node optimization

The distance between each gateway node and the fixed node is calculated. According to the distance, whether the gateway node has stability can be judged. The formula of the weighted average method is as follows: (3)${\displaystyle G_a=\alpha H_{na}+\beta {\mathrm{LQ}}_{na}+\theta E_{na}+\varepsilon {\mathrm{LD}}_{na}+\delta {\mathrm{SY}}_{na}.}$

Where Eq. (3) shows that: G_a represents the node selection factor, which is used to describe the selected node a that can be regarded as a comprehensive measure of a gateway; node n to gateway node α is defined as H_na, the weight is expressed as α; node n to gateway node α, the link quality is defined as LQ_na, and the weight is expressed as β; It should be noted that the overall performance of a node is inversely proportional to its selection factor, implying that a smaller selection factor indicates a higher level of comprehensive performance.

The stability of the primary gateway nodes is contingent upon their associated main gateway nodes and the distance from node n. A greater degree of stability in the fluctuation of distance signifies a superior level of stability in the gateway node. Utilizing the secondary gateway node as the gateway facilitates the expeditious and precise distribution of multimedia teaching resources. (4)${\displaystyle {\mathrm{SY}}_{na}=D_t^{n,a}=\frac{1}{n}\sum _{m=0}^{n-1}{\left\{d_{t-m}^{n,a}-E\left(d_t^{n,a}\right)\right\}}^2.}$

The average distance between node n and gateway node a at time t is shown in Eq. (5): (5)${\displaystyle E\left(d_t^{n,a}\right)=\frac{1}{n}\sum _{m=0}^{n-1}{d}_{t-m}^{n,a}.}$

The selection process entails identifying the value with the most minimal variance, indicating the optimal gateway node. This node is then designated as the gateway exit, enabling access to the external network through this node. By employing the node selected at this juncture, the intelligent sharing of multimedia teaching resources and the attainment of enhanced speed and precision in multimedia sharing across multiple nodes can be accomplished within the network.