An Innovative Method for the Assessment of Soil Moisture Using Auto Encoders in Comparison with Linear Discriminant Analysis for Better Accuracy

The main goal of the study is to evaluate the efficacy of auto encoders in comparison to linear discriminant analysis (LDA), with a particular emphasis on understanding soil moisture content. Agricultural productivity is improved by precise irrigation scheduling, optimal crop water use efficiency, decreased water waste, and a decreased risk of crop stress and yield losses from both water scarcity and waterlogging. The soil moisture dataset, which contains extensive soil data such dry soil, wet soil, and varied moisture conditions, makes up the materials used in this study. Two groups, each with 400 samples, form the basis of the study's framework. Samples from Group 2 are processed using the Linear Discriminant Analysis, while samples from Group 1 are handled using Auto encoders. A comprehensive and fair comparison of the two methods is made possible by this structure, which has an 800 sample size overall. In order to ensure sufficient power for the comparison study, sample size is calculated by statistical analysis. Using a 95% confidence interval and alpha (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\alpha$</tex>) and beta (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\beta$</tex>) values of 0.05 and 0.2, respectively, this estimate is based on a G-power of 80 %.These specifications were made to ensure an accurate and statistically significant assessment of the algorithm's performance. Accuracy is the primary measure used to compare the two models' performance. The accuracy of the Linear Discriminant Analysis was 65.07 %, whereas the Auto encoder had a better accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{7 5. 3 4 \%}$</tex>. A two-tailed significance test was used to further examine this notable performance difference. A <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p}$</tex>-value of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{0. 0 0 7}$</tex>, less than the significance limit of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0.05(\mathrm{P}&lt;0.05)$</tex>, was obtained from the test. This study conclusively suggests that auto encoders, as compared to linear discriminant analysis, provide a more precise and dependable technique for detecting soil moisture.