Time Series Modeling for Drought Stress Propagation in Plants using Hyperspectral Imagery

Choudhury, Sruti Das; Saha, Soumik; Samal, Ashok; Mazis, Anastasios; Awada, Tala

doi:10.22541/au.166733729.99500258/v1

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, diverse machine learning models have proven effective in accurately forecasting and refining plant tissue culture procedures. These models have been applied in various investigations, including in vitro mutagenesis, micropropagation, regeneration studies, plant system biology, in vitro organogenesis, stress physiology, and salt stress [28][29][30][31][32][33][34][35][36][37][38][39][40][41][42]. Only a few studies have used machine learning models to examine drought stress responses.…”

Section: Introductionmentioning

confidence: 99%

“…Gupta et al [33] developed an automation model for water stress detection in wheat using pre-processing and canopy segmentation methods, with the random forest algorithm achieving high accuracy. Das Choudhury et al [34] introduced HyperStressPropagateNet, a deep neural network for analyzing drought stress propagation in plants through hyperspectral imagery, showing a strong correlation with soil water content. Tahmasebi et al [35] applied a meta-analysis and machine learning to identify drought-responsive genes in Populus, revealing significant transcriptional variations and potential markers for breeding programs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning Offers Insights into the Impact of In Vitro Drought Stress on Strawberry Cultivars

Şimşek

2024

Agriculture

View full text Add to dashboard Cite

This study aimed to assess the susceptibility of three strawberry cultivars (“Festival”, “Fortuna”, and “Rubygem”) to drought stress induced by varying polyethylene glycol (PEG) concentrations in the culture medium. Plantlets were cultivated on a solid medium supplemented with 1 mg/L BAP, and PEG concentrations (0, 2, 4, and 6 mg/L) were introduced to simulate drought stress. Morphological changes were observed, and morphometric analysis was conducted. Additionally, artificial neural network (ANN) analysis and machine learning approaches were integrated into this study. The results showed significant effects of PEG concentrations on plant height and multiplication coefficients, highlighting genotype-specific responses. This study employed various machine learning models, with random forest consistently demonstrating superior performance. Our findings revealed the random forest model outperformed others with a remarkable global diagnostic accuracy of 91.164%, indicating its superior capability in detecting and predicting water stress effects in strawberries. Specifically, the RF model excelled in predicting root length and the number of roots for “Festival” and “Fortuna” cultivars, demonstrating its reliability across different genetic backgrounds. Meanwhile, for the “Rubygem” cultivar, the multi-layer perceptron (MLP) and Gaussian process (GP) models showed particular strengths in predicting proliferation and plant height, respectively. These findings highlight the potential of ML models, particularly RF, to enhance agricultural breeding and cultivation strategies through accurate phenotypic predictions, suggesting a promising direction for future research to improve these predictions further. This research contributes to understanding strawberry responses to drought stress and emphasizes the potential of machine learning in predicting plant characteristics.

show abstract