Логический подход к построению модели машинного обучения для оценки устойчивого развития горных территорий

Abstract

Introduction. There are a significant number of mountainous areas on the territory of Russia. The Russian Federation has extensive mountain systems, including the Altai Territory, the Caucasus region, the Ural Mountains, the Sayans, the Baikal region, Kamchatka, Chukotka and others. The exact number of mountainous areas is difficult to determine, since this concept can vary depending on the classification criteria and the boundaries of mountainous regions. This paper proposes an approach to constructing a machine learning model to determine the level of sustainable development of mountain regions. The sustainability of regional development is considered at three levels: environmental, social and economic. To build a model that can assess the level of sustainability of a region, various input data are used, including climatic conditions, geographical characteristics, socio-economic indicators and other variables associated with mountainous regions. After data collection, the most relevant features related to the sustainable development of the region are selected using correlation analysis, feature selection methods or expert knowledge. Next, a model is built based on logical methods of data analysis and processing. The choice of model is due to the fact that logical methods are better at handling small amounts of data and may be less prone to overfitting; they can ignore outliers or noisy data that negatively affect other machine learning methods. Logical methods can also be useful when working with unstructured data, such as texts or images. The work uses logical rules and patterns to extract information and classify data. The work proposes a software package that provides the ability to assess the sustainability of mountain areas based on a database, as well as create and update a knowledge system to improve the accuracy and adaptability of the sustainability assessment process. The goal of the work is to create a model that can help in assessing and predicting the sustainability of mountain areas, as well as in understanding the factors influencing this sustainability. This can help improve the planning and management of mountain regions to achieve more sustainable development and preserve their environmental and socio-economic integrity. Research methods. After data collection, the most relevant features related to the sustainable development of the region are selected using correlation analysis, feature selection methods or expert knowledge. Next, a model is built based on logical methods of data analysis and processing. The choice of model is due to the fact that logical methods are better at handling small amounts of data and may be less prone to overfitting; they can ignore outliers or noisy data that negatively affect other machine learning methods. Research result. A machine learning logic model that demonstrates its effectiveness in identifying and identifying the main factors influencing the sustainable development of a region.