Research on Management and Students’ Innovation Ability Training Method Based on Swarm Optimization Algorithm

Опубликована: Фев. 23, 2024

The research of training methods is critical in the management college education and students' innovation ability, however it has an issue with erroneous performance positioning. typical Ant colony algorithm unable to address ability result insufficient. As a result, Bee optimization algorithm-based tresearch on student method provided, assessed. To begin, swarm intelligence theory used discover influencing elements, indicators are split based method's needs decrease interference factors methods. then create scheme, outcomes thoroughly examined. MATLAB simulation results reveal that, under particular evaluation conditions, outperforms standard terms accuracy time variables.

Язык: Английский

---

Design of an Automatic Scoring System for Text Translation Information under XML Structure

Опубликована: Фев. 23, 2024

In recent years, technologies such as text segmentation, data mining, and machine learning have become the main research directions in field of document translation modern society. Especially, analysis has an important development trend information retrieval. Based on thorough summary, this study adopted XML (Extensible Markup Language) keyword extraction, feature selection, classification algorithms to design implement a simple, efficient, low-cost, executable recommendation system framework. This method can significantly improve understanding ability classical Chinese, thereby improving quality article. The also modular structure, making extraction process simpler faster. Afterwards, article conducted performance test automatic scoring for information. results showed that accuracy word was stable at 100%, while other translations above 85%; still performed stably translating words, but there were shortcomings consistency long paragraphs, with minimum rate 80%; fluency could reach up 100% 72%. aims automatically structured text, providing effective way evaluate compare different results.

Язык: Английский

---

English Classroom Teaching Based on Improved Apriori Algorithm

Опубликована: Фев. 23, 2024

Classroom teaching is the primary organizational formation of school and basic way obtaining student's creativity quality-oriented education. English teachers work from various backgrounds which include secondary schools, universities, colleges accommodating their techniques to suit different learning requirements students. However, only by maximizing capability in information age could understand students better fulfill student environment a network. In this research, Improved Apriori Algorithm (IAA) proposed for classroom quality. The IAA improved utilizing Boolean matrix row-column compression minimizes transaction database scanning time Trie tree employed increase search process. Initially, questionnaire sample data established evaluate teacher criteria tasks. preprocessing performed employing filtering then evaluation filtered text split into short sentences. Then, Word2 vector approach extract features. Finally, technique quality teaching. When compared with (AA), achieves innovation ability satisfaction 90% 95% respectively.

Язык: Английский

---

Application of Artificial Intelligence Technology in Enterprise Financial Audit

Опубликована: Фев. 23, 2024

In the era of big data, imperative for financial audits in enterprises to evolve towards data-centric and practical applications has become increasingly apparent. As technology progresses, it necessitates that adapt reform alignment with dynamic needs business development. This paper provides an insight into interplay between artificial intelligence (AI) enterprise finance, underpinned by a thorough analysis extensive existing data showcase significant role AI enhancing efficiency effectiveness audits. Currently, audit sector's reliance on substantial human resource investment technological approaches reached plateau, where further investments do not yield proportional quality breakthroughs. The argues industry-wide shift leveraging new technologies, particularly AI, catalyze transition more networked, digital, intelligent management system. is envisioned foster development paradigm industry, characterized virtuous cycle capital investment, innovation, income growth, thereby setting trajectory industry advancement.

Язык: Английский

---

Image Recognition Algorithm of UAV Inspection under Machine Vision

Опубликована: Фев. 23, 2024

UAVs have the advantages of efficient and automated inspection in life, important application value industry, construction, energy other fields. In this paper, an improved image recognition algorithm based on machine vision is proposed to solve problems existing UAV inspection. Through use computer technology, paper analyzes processes images captured by unmanned aerial vehicles complete automatic defect detection objects. A system composed preprocessing, feature extraction, object classification, also proposed. Using large-scale drone patrol data, method evaluated from three aspects: accuracy, F1 speed. After testing, prediction accuracy can reach 89% 96%. The has made significant improvements target detection. improvement score indicates that identify target.

Язык: Английский

---

Customer Churn Prediction in the Telecom Sector with Machine Learning and Adaptive k-Means Cluster using Imbalance Data

Опубликована: Март 15, 2024

A crucial aspect of maintaining a customer-oriented business in the telecommunications sector with machine learning (ML) is understanding reasons and factors that lead to customer churn. However, dataset difficult by noise, misclassifications, duplicated data, imbalanced information complicating process identifying ways split based on events. The under-sampling method address imbalance data decreasing quantity majority class, thereby achieving balanced dataset. This method, utilizing Adaptive k-means clustering, directly determines reduction quality each class. datasets are combined class labels generate new used for classification SVM algorithm, which distinguishes between churn non-churn prediction telecom sector. algorithm can perform linear non-linear kernel function control overfitting improve generalization handle high-dimensional data. obtained results demonstrate proposed achieves better accuracy 95.70%, precision 56.01%, Recall 55.05%, F1-score 96.05% These ensure superior detection performance compared other existing methods, such as Random Forest (RF), Decision Tree (DT) K-Nearest Neighbour (KNN).

Язык: Английский

---

Optimized Scheduling Strategy of Multi-Micro-Grid Integrated Energy System Considering the Efficient Utilization of Energy Cascade

Опубликована: Март 15, 2024

The implementation of the two-carbon target "carbon peak" and neutrality". increase comprehensive energy micro-grid will form a cluster, namely multi-micro-grid integrated system. Compared with single system can greatly improve cost grid. How to further overall use efficiency is still key point research. In this paper, an optimal scheduling model established for multi-micro grid including electric interaction considering efficient utilization cascade. Firstly, architecture quality coefficient method are detailed, differences between calculation thermal compared in detail. Subsequently, economy constructed. Matalb Yalmip used optimization modeling, Gurobi solver solve problem, which effectively avoids problem too long time heuristic algorithm. rationality effectiveness proposed scheme demonstrated through example analysis, not only improves efficiency, but also ensures high economy, making operation more line requirements target.

Язык: Английский

---

Vine Disease Detection UAV Multi Spectral Image using Segnet and Mobilenet Method

Опубликована: Март 15, 2024

Deep Learning-based vine disease detection has garnered significant attention from the community, particularly with utilization of UAV multispectral images for grapevine detection. However, identifying diseases in various crop and horticultural conditions remains a complex challenge, especially under mobile edge computing conditions. The process makes use PlantVillage dataset, which includes unlabelled data. Data normalization is performed UAVs are involved data capture, while SegNet architecture utilized segmentation. This enables separation healthy unhealthy vines Subsequently, classification using MobileNetV2, layers split to detect all combined spectral larger image sizes, greater than 32 × 32, resulting better performance. proposed method achieves high performance, an accuracy achieve 99.50%, precision at 99.42%, recall 99.39%, mean average (MAP) 99.20%. These metrics compared existing methods such as Convolutional Neural Network (DCNN) Inception V2.

Язык: Английский

---

Computer Information Extraction Algorithm Based on English Corpus

Опубликована: Март 15, 2024

Computer information extraction algorithms automatically extract valuable from a large amount of unstructured and non-standard text, such as person name, location, attribution, organizational structure, date, etc. This article focused on the research computer based English corpora. Firstly, this conducted simulation performance testing experiments an corpus to evaluate algorithm proposed in paper. In test paper, it can be seen that when dataset was 500, total precision 91%, while neural network traditional were 81% 70%, respectively, which lower than algorithm. Moreover, precision, recall F-value higher those outperformed several evaluation indicators, F-value. addition, more solvable dealing with problems linguistic diversity complexity grammatical structures. The is helpful promote development advancement algorithms. paper has provided ideas for exploring advanced deep learning models integrating variety

Язык: Английский

---

English Deep Learning Model Based on Improved SM-2 Algorithm

Опубликована: Март 15, 2024

In recent years, with the increasing level of global integration and English internationalization, demand for learning has also grown rapidly. However, current deep is relatively weak. classroom, teaching model that usually teacher-centered, student passive learning, knowledge-oriented still very common. Students mechanically read imitate in seemingly seamless processes, although sometimes there are performances, group cooperation, other forms auxiliary teaching. few students can question or confuse learned content through pre-thinking self-understanding, they unable to use have solve some problems real life. this mode, students' only superficial cannot delve into English, which hinders improvement their overall abilities. Therefore, article optimized by improving SM-2 algorithm (elliptic curve public key cryptography algorithm), so deeply learn therefore easily pass TEM4 (Test Majors-Band 4).

Язык: Английский

---