From the perspective of artificial intelligence (AI), this paper explores the application and impact of cluster analysis in the criticism of narrative ethics in Chinese new century literature. Utilizing AI paper processing technology, a large amount of literary text data is quickly obtained and processed, and a knowledge map of narrative literary works is constructed. Meanwhile, a clustering algorithm is used to divide the keywords of literary works into cluster classes to improve the efficiency of rapid literary analysis. The regression model is used to evaluate the effect of the cluster analysis method in the AI perspective on the ethical criticism of literary narratives. The accuracy, recall, and F1 value of the two AI techniques selected in this paper in the classification of literary text themes, keywords, and emotions are 85% to 90%, which is higher than the comparison methods, and combined with the clustering algorithm, the keyword categories of the literary text can be obtained quickly and precisely. In addition, by constructing a knowledge graph, this paper can help users grasp the character relationships in literary texts more clearly and assist in ethical criticism. The investigators are highly satisfied with the method of this paper, the average rating of each dimension is between 4.09 and 4.7, and the method has a significant contribution to the effect of ethical criticism of literary narratives.

Generative artificial intelligence, as a new technology paradigm, has received more and more attention for its powerful generative ability and wide application prospects. Especially in automated control systems, the application of technology based on generative artificial intelligence is gradually becoming a hot spot of research. In this paper, the generative AI automation control system is divided into four levels: input layer, processing layer, instruction generation and control execution layer, and combined with dual encoders, the attention model of multilingual to semantic expression is constructed. Two-dimensional variables are selected to construct a fuzzy PID control system to realize automation control for generative AI system. Comparing the control effects of fuzzy control PID and classical PID, the average errors of the two systems are 1= , 2= respectively. The maximum overshoot and rise time are 9% and 0.08 s, 5% and 0.04 s. The fuzzy PID control effect is more accurate, and at the same time improves the dynamic performance of the system. Analyze the implementation effect on the innovative service application of generative artificial intelligence. Comparing the overall recognition effect of the control system B proposed in this paper, and the two systems with reference to A, their overall recognition effect indexes are 0.94755 and 0.87211, respectively, and the fuzzy PID control system plays an auxiliary enhancement role in the contextual feature recognition of translation services in the intelligent library.

The assessment of economic quality is of great significance in grasping the state of national economic development at a macro level. This paper focuses on exploring the assessment methods of economic quality and introducing deep learning models to improve the shortcomings of the traditional economic quality assessment in the assessment process. The economic quality assessment system is constructed from five dimensions, including economic vitality, and the MIV indicator values are improved by combining set-pair analysis and generalized regression neural network, so as to realize the automatic screening of economic quality evaluation indicators. According to the screening results of the indicators, the hierarchical analysis method is used to assign weights to the indicators, and the comprehensive index of economic quality is measured based on the results of the assignment.From 2012 to 2022, the economic quality of the 30 provinces in China shows an upward trend as a whole, and the comprehensive index of economic quality in 2022 is 0.90, which is an increase of 52.54% compared with that in 2012. The assessment results are consistent with the actual results, indicating that the method of this paper can effectively complete the measurement and assessment of the economic quality index, which is important for the study of economic quality.

The article builds a simulation system based on human physiological parameters, collects human physiological data through the human physiological model, and simulates human physiological signals. The load adaptability of trainers to aerobics training was explored by studying the changes in the SI values of T-lymphocytes of the subjects’ bodies during aerobics training. SPSS and independent samples t-test were used to analyze the exercise data of the experimental group and the control group, so as to verify whether the aerobics training has a good exercise effect. At Week 0, the SI values of T lymphocytes in the immediate post-exercise group and the 3-hour recovery group after exercise were 0.88, 0.61 and 0.70, respectively. In Week 2, it dropped to 0.34 and 0.49, respectively. At Week 6, the SI values of lymphocytes in the two groups were 0.60 and 0.30, respectively. The SI values of T lymphocytes in Week 0, Week 2, Week 4 and Week 6 in the quiet group were 0.88, 0.48, 0.80 and 0.50, respectively. Before the experiment, there was no significant difference between the experimental group and the control group in terms of exercise effect, and after the experiment, a significant difference was produced, and the exercise effect of the experimental group far exceeded that of the control group. The experimental group’s exercise effect improved by 6.43, 5.13, 6.91, 6.38, and 5.80 points on each of the five dimensions, a significant difference. The control group, on the other hand, remained essentially unchanged.

In today’s rapid development of information technology and big data technology, consumer behavior is undergoing a profound transformation. This study focuses on the decision-making stage of consumer journey, selects indicators based on webpage click stream data, improves the K-means algorithm, and realizes the identification of consumer journey nodes using the binary K-means algorithm. Based on the review recommendation scenario, from the perspective of consumer decision-making journey, we introduce the “attention-attitude-understanding-purchase intention” stage-based decision-making model, apply it to the model design of deep learning, and combine the attention mechanism and co-attention mechanism to propose a product recommendation method based on online reviews. The results show that consumers in clusters 1-4 are in the consumer journey nodes of attention, understanding, attitude, and purchase intention, respectively. The product recommendation model exhibits better recommendation accuracy and time efficiency, with accuracy improved by 18.72%~67.12% and time reduced by 8.39%~62.03% over the comparison method. This paper realizes the innovation of deep learning method with the support of consumer behavior theory, and improves the methodological technical support for accurate online marketing strategy.

The pattern design of clothing appearance is one of the important links in clothing design, which makes an important contribution to the overall aesthetics and sales of clothing. As a product of computer technology, the development and application of graphic processing technology has been extended to various industries and fields of society, especially in the field of design with more extensive use. However, the current clothing pattern design is still too dependent on the designer, so this paper is based on pattern processing, combined with fractal algorithm and genetic algorithm to build a pattern generation algorithm for clothing pattern. And the quality of the generated pattern is optimized based on the anti-alignment algorithm, so as to improve the overall quality of the generated pattern. After testing, the real-time generation speed of the pattern generation algorithms for clothing patterns in this paper is greater than 15FPS, and from the subjective and objective points of view, the generated patterns have good quality to meet the needs of use. After the anti-alignment optimization of this paper’s algorithm in different error intervals in the number of pixels accounted for the percentage of screen pixels are the highest, are more than 99%, to further validate the optimization effect of this paper’s method. Finally, in the evaluation of the use of the algorithm, the testers have a high degree of satisfaction with the dimensions of this paper’s algorithm, respectively, 4.04, 3.98, 4.21 and 4.11, which shows that this paper’s algorithm can satisfy the practical needs and can realize the intelligent generation of clothing pattern design.

The legal positioning of blockchain technology applied to evidence and its attributes are the basis for its evidence review and rule design. This paper starts from analyzing the evidence attributes of blockchain electronic data, combines relevant regulations and judicial interpretations, and clarifies the legal effect of blockchain electronic data. Combined with the judicial application of blockchain evidence at home and abroad, it points out the specialized review rules of blockchain evidence. Obtain the blockchain access evidence process, and propose the block file storage method based on RS code as well as the decryption outsourcing attribute-based encryption scheme with the same sub-policy to improve the CP-ABE encryption scheme. Explore the rules for blockchain deposits and clarify the rules and institutional value of blockchain deposits for admissibility. Analyze the theoretical and practical operational performance of the improved attribute-based encryption algorithm. Optimize the evidence storage capacity of blockchain, and analyze the performance of the blockchain technology scheme designed in this paper in the intelligent review of access evidence. In the forensic scenario run by the algorithm in this paper, the stored evidence data is reduced by 1417 characters, the transaction response time is shortened by 175.361ms on average, and the block size is reduced by about 4 times. It proves that the blockchain algorithm scheme proposed in this paper can shrink the cost of depositing evidence, reduce the time of depositing evidence, and improve the efficiency of depositing evidence in the public security forensic system.

As the material foundation of language, speech is the basis for mastering language skills and capturing language information, and English learning must begin with the correct mastery of spoken language. Therefore, spoken language teaching occupies a rather important position in English teaching. In this study, we extract various features such as time-domain features and frequency-domain features from English spoken audio signals, use fuzzy logic inference model to represent each audio feature mapping as an affiliation function, and then optimize the parameters of the affiliation function by using adaptive neuro-fuzzy inference system, and solve the affiliation function to get the result of speech matching by the center of gravity method. Subsequently, a speech evaluation system is designed based on the speech matching model to assist intelligent spoken language teaching. The results of teaching practice show that students in the experimental class using the voice assessment system as a learning aid are significantly better than the control class in terms of speaking skills and learning attitudes (P<0.05). Through real-time feedback and personalized practice, the voice assessment system enables students to correct pronunciation errors immediately and gradually improve their speaking fluency and accuracy. It can also improve students' self-efficacy and learning motivation. This study confirms the effectiveness of the fuzzy logic-based audio classification and speech matching model in improving students' spoken English proficiency and reveals its potential for wide application in future spoken English education.

Existing natural language generation models often face the problems of context loss and incoherent responses when dealing with multi-round dialogs. In this paper, a multi-round dialog system based on Transformer architecture is constructed, and an intention recognition algorithm is used to form a technical support for the construction of multi-round dialog system. And the attention adapter is introduced into the natural language generation module in the system, which utilizes contextual features to improve the performance of the natural language generation model. Semantic slot extraction experiments are carried out on the ATIS dataset, and the F1 values of the semantic slot extraction task and intention recognition task of the BERT multi-task natural language generation model with the addition of the attention mechanism are improved by 0.64% and 0.15%, respectively. The multi-round dialog system designed in this paper has a perplexity of 18.33, and the BLEU metrics are higher on orders 1-4 compared to other models. Manually evaluated in terms of syntactic semantic coherence, relevance, and information content, the system performs better. It shows that the natural language generation model incorporating the attention mechanism can effectively improve the application effect of the multi-round dialog system.

The continuous development of data analysis technology in the era of big data provides new methods for the analysis of college students’ ideological dynamic data, and also provides new ideas for the scientific construction of ideological and political education disciplines in colleges and universities. In this paper, based on the word frequency analysis, set up the keyword context, collect the keywords of the ideological dynamics of party members and students, extract and examine the feature vectors in them, put all the keyword feature vectors, form the keyword feature vector set, use the method of keyword vector research, carry out the descriptive and differential analysis of the survey data of the ideological dynamics of the party members’ development in the management of colleges and universities, and construct the party members’ ideological dynamics management mechanism based on the existing problems , analyzing the effect of management decisions in colleges and universities. The kurtosis values of the four indicators of party members’ political thoughts, learning thoughts, innovative thoughts and consumption thoughts are -1.4685, -0.4496, -0.9871 and -1.5614 respectively, which are on the low side, indicating that the performance of the respondents is more unified and concentrated in these four indicators, and laying a feasible foundation for the subsequent relevant analysis. In campus adaptation, self-efficacy, and life satisfaction, the number of party students who performed agreeably, correctly, and positively were 116, 202, and 142, respectively, and after the reform of college management decisions, the students’ performance in these three aspects changed.