DEVELOPMENT OF A CLOUD SERVICE FOR AUTOMATIC ANALYSIS OF TEXT DOCUMENTS OF DISTANCE LEARNING SYSTEMS - Scientific conference
Main page
UA PL EN

Congratulation from Internet Conference!

Hello

Рік заснування видання - 2011

DEVELOPMENT OF A CLOUD SERVICE FOR AUTOMATIC ANALYSIS OF TEXT DOCUMENTS OF DISTANCE LEARNING SYSTEMS

15.02.2026 22:37

[1. Information systems and technologies]

Author: Oleksandr Viunenko, Candidate of Economic Sciences, Associate Professor of Cybernetics and Informatics Department, Sumy National Agrarian University, Sumy



The growth in the volume and complexity of text data in various industries highlights the urgent need to implement effective and accurate document analysis methods. Traditional approaches to text analysis are often insufficient for processing the modern scale and complexity of text documents. In contrast, the integration of deep neural networks and cloud services shows promise in the field of document analysis automation. This approach allows users to obtain samples, generate explanations, and engage in interactive dialogue with neural network models. The development of a web service for automated text analysis meets this need by providing a powerful tool for improving productivity and knowledge discovery in various fields. The main problem lies in the complexity and inaccessibility of existing text document analysis services, such as Google Natural Language AI, Amazon Comprehend, and Lexalytics. These services, although powerful, are designed for developers and require a certain level of technical knowledge to use effectively. In addition, these services do not have a convenient interactive interface, which makes it difficult for ordinary users to effectively analyse text documents.

To solve this problem, it is necessary to develop a convenient web service that uses the capabilities of deep neural networks to analyse and explain text documents in interactive mode. This service should offer a chat interface that allows users to interact with the system in a natural, conversational manner, asking questions and receiving answers in real time. This will make the text analysis process more engaging and intuitive, removing the barriers often associated with more technical systems. The integration of modern deep neural network models, such as OpenAI's GPT-4 model, into the web service is an important aspect of enabling automated analysis and interpretation of text documents. This process involves several key steps to ensure effective interaction between the web service and the neural network model, as well as the generation of accurate and contextually relevant responses. Other models, such as BERT, may also be considered depending on specific requirements [1]. As with any external API integration, it is necessary to have error handling mechanisms in place to address issues such as network failures, API rate limits, or unexpected model errors. Implementing retry strategies and fallback mechanisms can help ensure service reliability. 

Effective data management is crucial to ensuring the smooth operation of a web service, especially when processing user documents, requests, and responses. Data management encompasses several key components, including storage, search, and integration with external services. Use of third-party services to store PDF documents uploaded by users. After a document is uploaded, the service generates a unique URL, which is then stored along with additional metadata in a relational database. This metadata includes information such as the user ID, document title, upload timestamp, and any relevant tags or categories associated with the document.

Semantic understanding is a key aspect of the web service, ensuring accurate interpretation of the meanings of text documents and user queries. The web service uses the OpenAI Embeddings API to convert text documents and user queries into semantic vectors [1]. A semantic vector is a dense, high-dimensional representation of text that reflects the semantic relationships between words and phrases. By converting text into vector space, the web service can perform semantic operations such as similarity comparisons and contextual analysis.

When a user uploads a document, its content is processed to extract textual information. This text is then passed through the Embeddings model to create a semantic vector representation. The vector encapsulates the semantic meaning of the document, reflecting its key concepts, themes, and relationships between words.

Semantic search methods are used to match the semantic vectors of user queries with the content of documents, involving the calculation of similarity scores between vectors using methods such as cosine similarity. Documents with vectors most similar to the query vector are found and considered as potential context for generating responses. Once relevant documents have been identified, their semantic vectors are used to provide context to deep neural network models. By incorporating this context into the analysis process, models can generate more accurate and contextually relevant responses to user queries [2]. Semantic understanding is the basis for a web service's ability to analyse text documents and provide meaningful explanations. Using the Embeddings model and semantic search methods, the service can effectively bridge the semantic gap between user queries and document content, enabling more accurate and in-depth analysis. Effectively solving these problems allows for the development of a reliable cloud-based educational web service capable of providing automated analysis and explanations of text documents in an interactive and in-depth manner.

Text document analysis using deep neural networks is a powerful tool for extracting meaningful information from large volumes of training text data. Text document analysis methods include information retrieval, text summarization, topic modelling, sentiment analysis, named entity recognition, and text classification. DNNs are used to perform these tasks with high accuracy, scalability, and versatility. The creation of such web services for analysing educational text documents also solves the problem of forming individual learning trajectories for students. They can be specifically designed to meet the specific needs of a general audience and enable users not only to obtain text analysis, but also to interact with the data in the most convenient and effective way.

References

1. Horev R. BERT Explained: State of the art language model for NLP. Towards Data Science. URL: https://towardsdatascience.com/bert-explained-state-of-the-art-language-model-for-nlp-f8b21a9b6270 (date of acces: 20.01.2026).

2. Tripathi R. What are Vector Embeddings. Pinecone. URL: https://www.pinecone.io/learn/vector-embeddings/ (date of acces: 20.01.2026).



Creative Commons Attribution Ця робота ліцензується відповідно до Creative Commons Attribution 4.0 International License
допомога Знайшли помилку? Виділіть помилковий текст мишкою і натисніть Ctrl + Enter

Another articles in this section

Сonferences

Conference 2026

Conference 2025

Conference 2024

Conference 2023

Conference 2022

Conference 2021



Міжнародна інтернет-конференція з економіки, інформаційних систем і технологій, психології та педагогіки

Наукова спільнота - інтернет конференції

:: LEX-LINE :: Юридична лінія

Інформаційне суспільство: технологічні, економічні та технічні аспекти становлення