Explore the vast range of titles available.

In this paper, we use big data analysis techniques combined with the TF-IDF algorithm to weigh the frequently occurring word frequency vectors in text and reduce the document length to obtain keywords without destroying the original text feature information. The similarity of text features is combined with a Bayesian algorithm for label classification to facilitate data query and indexing. The results show that the running time of the system is kept around 14s, the recall and accuracy can be close to about 75% and 72% on average, and the number of keywords can reach 5971 with an F1 value of 0.9, which proves the effectiveness of the artificial intelligence legal regulation system based on big data analysis.

Due to the slow processing speed of text topic clustering in stand-alone architecture under the background of big data, this paper takes news text as the research object and proposes LDA text topic clustering algorithm based on Spark big data platform. Since the TF-IDF (term frequency-inverse document frequency) algorithm under Spark is irreversible to word mapping, the mapped words indexes cannot be traced back to the original words. In this paper, an optimized method is proposed that TF-IDF under Spark to ensure the text words can be restored. Firstly, the text feature is extracted by the TF-IDF algorithm combined CountVectorizer proposed in this paper, and then the features are inputted to the LDA (Latent Dirichlet Allocation) topic model for training. Finally, the text topic clustering is obtained. Experimental results show that for large data samples, the processing speed of LDA topic model clustering has been improved based Spark. At the same time, compared with the LDA topic model based on word frequency input, the model proposed in this paper has a reduction of perplexity.

With the development of information technology and the change of people’s education concept, personalized learning is getting more and more attention. The traditional classroom is difficult to realize the need of personalized English reading comprehension material recommendation, this paper designs a fusion English reading comprehension material recommendation system based on collaborative filtering algorithm and improved text similarity algorithm. Aiming at the problem that the traditional text similarity algorithm ignores the user’s differentiated attention to the information of English reading comprehension materials in text matching, the Inter-TF-IDF algorithm, which integrates the calculation of concentration and dispersion, is proposed. Combining this algorithm and the user-based collaborative filtering algorithm, the fusion recommendation system of this paper is proposed. The system utilizes the collaborative filtering algorithm to get the preliminary recommendation results of English reading comprehension materials, and then utilizes the improved Inter-TF-IDF algorithm to calculate the similarity between the text of the English reading comprehension materials and the text browsed by the user, and selects the English reading comprehension materials with a high degree of similarity as the final recommendation results. The overall recommendation accuracy of the system in this paper is maintained at a high level of 0.82-0.95. In the actual application effect, it significantly improves the English scores of students using the system, and obtains a high degree of satisfaction from the students. The recommendation system in this paper has a good promotion and application prospect in the field of English learning.

As consumption upgrades, China’s cultural and creative industry is experiencing rapid growth. However, user empathy remains lacking in some products. This paper explores a design method for these products based on empathy theory. It employs the TF-IDF algorithm to extract semantic features from product texts and uses the SVM algorithm for classification. Post-classification, the LDA theme model analyzes sentiment, integrating both visual and semantic models to enhance product design. An analysis of cultural and creative products using data analysis software reveals that approximately 79.5% of user comments exhibit positive sentiment, with an average sentiment score of 2.9218 and a peak score of 39.1363. This suggests strong positive emotional responses from nearly 80% of users. The proposed method effectively enhances user-product interaction and empathy.

Starting from the basic principles of talent cultivation mode, this paper proposes a reform strategy to optimize the talent cultivation mode of environmental design with the help of intelligent technology in order to improve the quality of talent cultivation and achieve the “first-class” level for the current talent cultivation mode of environmental design majors. The necessity of an intelligent database in environmental design teaching is analyzed, and the improved TF-IDF algorithm and automatic classification algorithm of resources are proposed to improve the search for teaching resources for environmental design majors. Then, based on the query, the teaching resource recommendation model (UBM) is proposed. An algorithmic comparison between the improved TF-IDF method and the UBM model is carried out to verify the effectiveness of the improved algorithmic model. Combining the traditional talent training model results with the intelligent technology-assisted talent training model results for independent samples of the T-value test. T-value test results show that the F-value is 0.317, the Sig-value is 0.782>0.05, and the significance corresponding to the ANOVA t-test value is 0.009<0.05. It indicates that there is a significant difference in the final examination results for environmental design in the classes of the experimental and control groups. It means that after one semester of intelligent technology-assisted teaching practice, the professional performance of environmental design students improves, which contributes to the cultivation of first-class professional talents. The use of an intelligent database can be one of the effective ways of cultivating talent.

The urbanization process is accelerating, the competition between cities is getting more and more intense, and the problem of urban image management is getting more and more attention from urban image researchers and city managers. Taking Nanchang City as the case study, this paper selects the city government, economy, tourism, and people image, as well as three data platforms: the official website, news positioning, and tourists’ diary, and crawls the comments of the four types of city image elements on the platforms, and carries out the data fusion and pre-processing to get the comment dataset. Subsequently, the TF-IDF algorithm is used to extract high-frequency words from the city image, and it is combined with the SnowNLP model and LDA theme analysis model to analyze and supplement the overall image of Nanchang City. Finally, after studying the mechanism of city image under data (media) fusion, the communication effect of Nanchang’s image is explored in terms of communication heat, recognition, and participation. Government image and tourism image aspects of the words appear most often, and the highest and lowest percentage of positive emotions are the city’s economic image and government image, respectively, which should integrate the existing resources, improve the level of service, and further create the city’s characteristics, to expand the city’s visibility and communication effect.

Construction accidents can lead to serious consequences. To reduce the occurrence of such accidents and strengthen the execution capabilities in on-site safety management, managers must analyze accident report texts in depth and extract valuable information from them. However, accident report texts are usually presented in unstructured or semi-structured forms; analyzing these texts manually requires a lot of time and effort, it is difficult to cope with the demand of analyzing a large number of accident texts, and the quality of key information extracted manually may be poor. Therefore, this study proposes a classification method based on natural language processing (NLP) technology. First, we developed a text classification model based on a convolutional neural network (CNN) that can automatically classify accident categories based on accident text features. Next, taking the classified fall accidents as an example, we extracted key information from accident narratives using the term frequency-inverse document frequency (TF-IDF) method and presented it visually using word clouds. The results show that the overall accuracy of the CNN model reaches 84%, which is better than the other three shallow machine-learning models. Then, eight key accident areas and three accident-prone operations were identified using the TF-IDF algorithm. This study can provide important guidance for project managers and can be used for on-site safety management to help prevent production safety accidents.

This paper proposes a research method of personalized recommendation of test questions for users by fully collecting and studying the user’s characteristic data in multiple dimensions, establishing a dynamic user portrait, and mining and analyzing the characteristic information of test questions in the test question bank. Firstly, collect the user’s data, calculate the user’s similarity by analyzing the user’s behavior sequence, and use the cosine similarity algorithm to calculate the text similarity of the test questions. Then, the tag system is established to analyze the user activity, and the TF-IDF algorithm is used to calculate the user’s proficiency in different keyword classification questions. Based on the above information, a hybrid recommendation mechanism is adopted to provide users with personalized test question recommendation services.

This study effectively associates the evaluation object tree with attribute nodes by constructing an innovative tree-structured emotion dictionary. Utilizing the TF-IDF algorithm, this paper meticulously classifies the emotional features in opera librettos and adds a list of inspirational word nodes. Further, a basic framework of weighted decision matrix is designed, and the data in the emotion score matrix is normalized by Softmax method, to derive the weight allocation coefficients, according to which the results of weight coefficient allocation of the opera libretto are calculated. In particular, for the adaptation of the famous opera Macbeth, this paper proposes three significant principles to ensure that the emotional Expression of the libretto matches the creative intent. The adapted libretto was subjected to emotion calculation and effect analysis, and the empirical study showed that in the first six minutes of the adaptation, there were two periods characterized by extremely significant emotional tension, in which the emotional tension exceeded 7.5 points in both cases. In the specific dynamic analysis, about 60% of the probability emotional value of clip 2 is located in the first quadrant, which coincides with the expected creative goal. In the correlation analysis of the script adaptation, the correlation between emotional attachment and the audience’s willingness to watch the play is as high as 0.8963, indicating that the adaptation is effective. This study not only innovates the method of emotion analysis of opera libretto in theory, but also provides practical guidance for opera libretto adaptation in practice, which helps to enhance the artistic infectivity of opera works and the audience’s viewing experience.

This paper utilizes the EA model to create and implement the top-layer architecture of education information systems for college students. Firstly, the design of the data layer is realized based on the improvement of the Solr rule and TF-IDF algorithm, the design of the business layer is realized from the user’s point of view, the design of the application layer is formed with the focus on the business link, and the design of integration-oriented software technology layer is realized. Then, it designed the practice of teaching informatization mode of college students’ civic and political education as an example. Finally, the performance test of Civic and Political Education Resources Retrieval and the practice effect of Civic and Political Education Teaching Informatization Mode are analyzed. The results show that the retrieval accuracy exceeds 0.9, the retrieval time range is within 1s, the experimental group’s worldview, outlook on life, patriotic consciousness, civic consciousness and the concept of the rule of law have improved by about 5 points in comparison with the general group, and the overall satisfaction with the informationization model of education and teaching is between 0.8 and 1.0.