Analysis of Forecasting Accuracy Depending on Various Hyperparameters of Neural Network Training. In Proceedings of the Computational Methods in Systems and Software : доклад, тезисы доклада

Описание

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: The 2024 International Conference on Computational Methods in Systems and Software (CoMeSySo 2024); Online; Online

Год издания: 2025

Идентификатор DOI: 10.1007/978-3-031-96798-6_9

Ключевые слова: neural networks, hyperparameters, Model Architecture, prediction accuracy, machine learning, data standardization, retraining

Аннотация: In this paper, a study was conducted on the accuracy of forecasting real estate prices using various neural network architectures. The results showed that the two-layer model with 32 neurons in each layer achieved the highest prediction accuracy (0.98), demonstrating the ability to capture complex data dependencies most effectivelyПоказать полностьюand avoid overfitting. This study highlights the importance of choosing the optimal neural network architecture and configuring hyperparameters to achieve high prediction accuracy. The optimal configuration must balance the complexity of the model and its ability to generalize.

Ссылки на полный текст

Издание

Журнал: Artificial Intelligence for System Oriented Design

Выпуск журнала: 1489-1

Номера страниц: 95-102

Место издания: Cham

Персоны

  • Kukartsev V.V. (RSAU-MAA named after K.A. Timiryazev)
  • Degtyareva K.V. (Bauman Moscow State Technical University)

Вхождение в базы данных