Afif (2025) Drought prediction data for IoT

Identification

• Journal: Mendeley Data
• Year: 2025
• Date: 2025-12-15
• Authors: Afif, Mahin Montasir
• DOI: 10.17632/vf3txm2c4y.1

Research Groups

• Mahin Montasir Afif (Contributor)

Short Summary

This paper presents a synthetic daily drought prediction dataset spanning two years, designed to simulate realistic soil and rainfall conditions for research in drought prediction and Internet of Things (IoT) applications.

Objective

• To create a synthetic daily dataset of soil and rainfall conditions, including a binary drought label, suitable for training machine learning and deep learning models for drought prediction in IoT contexts.

Study Configuration

• Spatial Scale: Simulated environmental conditions, not tied to a specific geographical location.
• Temporal Scale: 2 years (730 days) of daily measurements.

Methodology and Data

• Models used: The dataset itself was generated using seasonal sinusoidal patterns, random noise, and rule-based conditions for drought labeling. It is intended for training sequence-based models (e.g., GRU, LSTM) or classical machine learning models.
• Data sources: Synthetic data generated through simulation, not derived from real-world observations or reanalysis.

Main Results

• A synthetic dataset comprising 730 daily entries with features including soil moisture percentage (5–80%), soil wetness indicator (binary), daily rainfall percentage (0–100%), rain presence (binary), and a binary drought label.
• The dataset simulates realistic environmental variability through sinusoidal patterns and random noise, with drought labels derived from combined low soil moisture, absence of rain, and dry soil conditions.

Contributions

• Provides a novel, publicly available synthetic dataset specifically tailored for drought prediction research, particularly for IoT applications and machine learning model training.
• Offers a controlled environment for experimenting with real-time drought detection systems and educational purposes, overcoming limitations of real-world data availability or privacy.

Funding

• Not specified in the provided text.

Citation

@article{Afif2025Drought,
  author = {Afif, Mahin Montasir},
  title = {Drought prediction data for IoT},
  journal = {Mendeley Data},
  year = {2025},
  doi = {10.17632/vf3txm2c4y.1},
  url = {https://doi.org/10.17632/vf3txm2c4y.1}
}