Yield prediction for crops by gradient-based algorithms

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(8), P. e0291928 - e0291928

Published: Aug. 26, 2024

A timely and consistent assessment of crop yield will assist the farmers in improving their income, minimizing losses, deriving strategic plans agricultural commodities to adopt import-export policies. Crop predictions are one various challenges faced agriculture sector play a significant role planning decision-making. Machine learning algorithms provided enough belief proved ability predict yield. The selection most suitable is influenced by environmental factors such as temperature, soil fertility, water availability, quality, seasonal variations, well economic considerations stock preservation capabilities, market demand, purchasing power, prices. paper outlines framework used evaluate performance machine-learning for forecasting yields. models were based on range prime parameters including pesticides, rainfall average temperature. Results three machine algorithms, Categorical Boosting (CatBoost), Light Gradient-Boosting (LightGBM), eXtreme Gradient (XGBoost) compared found more accurate than other predicting RMSE R 2 values calculated compare predicted observed rice yields, resulting following values: CatBoost with 800 (0.24), LightGBM 737 (0.33), XGBoost 744 (0.31). Among these demonstrated highest precision achieving an accuracy rate 99.123%.

Language: Английский

Mechanisms of Rural Sustainable Development Driven by Land Use Restructuring: A Perspective of “Scale-Space” Interactions

Sustainability, Journal Year: 2023, Volume and Issue: 15(16), P. 12600 - 12600

Published: Aug. 20, 2023

Regional development issues can be reflected in land use and addressed through restructuring. It is commonly recognized that strengthening the role of small towns coordinated sustainable counties, towns, villages effective solving problems. However, relevant studies do not focus on county space lack scale extensions to reflect status towns. In this paper, an analytical framework developed some research hypotheses are proposed from perspective “Scale-Space” interaction. Then, a central China used as case study reveal mechanism rural restructuring by comparing changes structure different locations recent years. The finds under constraint basic principle “Requisition-Compensation Balance”, has reorganized spatial landscape pattern countryside extensive “demolition old” “construction new”. Among them, distant center, while improving their cohesion, give full play regional sub-centers, expected become important force driving urbanization future.

Language: Английский

Implications for conservation zoning in Teiči Strict Nature Reserve due to land use and cover change

Frontiers in Remote Sensing, Journal Year: 2024, Volume and Issue: 5

Published: Nov. 28, 2024

This study investigates land use and cover changes in the Teiči Strict Nature Reserve, Latvia, from 1982 to 2023 models potential future based on four legislative scenarios extending 2064. The research aims assess effectiveness of existing conservation zoning relation historical projected cover, addressing ongoing debate between fixed adaptable strategies. employs remote sensing imagery, geographical information system data, modeling methods analyze predict trends. results suggest substantial over past decades, including intensified urbanization, agricultural expansion, a decline peat bogs forested areas. Scenario projections indicate that continued could further challenge efficiency current strategies, with varying outcomes depending measures climate change impacts. concludes adaptive management variable are necessary address these dynamic preserve reserve’s ecological integrity. emphasize importance integrating predictive into planning improve flexibility sustainability protected

Language: Английский