Linear Multi-Task Learning for Predicting Soil Properties Using Field Spectroscopy

Abstract

Abstract:Field spectroscopy has been suggested to be an efficient method for predicting soil properties using quantitative mathematical models in a rapid and non-destructive manner. Traditional multivariate regression algorithms usually regard the modeling of each soil property as a single task, which means only one response variable is considered as the output during modeling. Therefore, these algorithms are less suitable for the prediction of several key soil properties with low concentrations or unobvious spectral absorption signals. In the current study, we investigated the performance of a linear multi-task learning (LMTL) algorithm based on a regularized dirty model for modeling and predicting several key soil properties using field spectroscopy (350-2500 nm) as an integrated approach. We tested seven key soil properties including available nitrogen (N), phosphorus (P) and potassium (K), pH, water content (WC), organic matter (OM), and electrical conductivity (EC) in drylands. The model performances of LMTL models were compared with the commonly used single-task algorithm of the partial least squares regression (PLS-R). Our results show that the LMTL models outperformed the PLS-R models with the advantage of shared features; the ratio of performance to deviation (RPD) values in the validation set improved by 10.24%, 4.93%, 25.77%, 11.76%, 6.74%, 53.13%, and 3.15% for N, P, K, pH, WC, OM, and EC, respectively. The best prediction was obtained for OM with RPD = 2.29, indicating high accuracy (RPD > 2). The prediction results of N, P, WC, and pH were categorized as of moderate accuracy (1.4 < RPD < 2), while K and EC were categorized as of poor accuracy (RPD < 1.4). However, the explanatory power of the LMTL models was moderate due to fewer features being selected by the regularization algorithm of the LMTL approach, which should be further studied in the soil spectral analysis. Our results highlight the use of LMTL in field spectroscopy analysis that can improve the generalization performance of regression models for predicting soil properties.

Combining Fractional Order Derivative and Spectral Variable Selection for Organic Matter Estimation of Homogeneous Soil Samples by VIS–NIR Spectroscopy

“…Because the field spectral reflectance is sensitive to interference with the atmospheric water absorption [34], reflectance spectra of these 91 fresh soil samples were acquired in the laboratory. An ASD FieldSpec 3 spectrometer with a spectral range of 350-2500 nm was applied to measure the soil VIS-NIR spectra.…”

Quantitative Remote Sensing of Land Surface Variables: Progress and Perspective

“…Qi et al [16] collected and analyzed soil spectra and soil samples from four watersheds in Israel and compared the relative performances of the linear multi-task learning (LMTL) algorithm and the partial least squares regression (PLS-R) single-task learning algorithm to derive different soil properties. The authors found that the multi-task algorithm had a slightly higher retrieval capability and was able to assess the organic matter with a good accuracy and a ratio of performance to deviation (RPD) higher than 2, the Ph, amount of nitrogen, phosphorus, and water content with a moderate accuracy (RPD between 1.4 and 2), and the amount of potassium and the electrical conductivity with a lower accuracy (RPD lower than 1.4).…”

“…Using field and laboratory-based spectroscopy, Liu et al [15] and Qi et al [16] demonstrated the potential of airborne and satellite hyperspectral measurements for large-scale soil property monitoring and mapping.…”

Untitled

“…The explanatory power of the independent variables to the dependent variables is achieved by calculating the VIP score. The independent variables are sequenced according to the explanatory power (Qi et al 2017). The VIP score for the j-th variable is given as:…”

“…Plenty of studies have demonstrated that spectral variable selection methods can not only reduce the complexity of calibration models, but also improve the model predictive performance (Hong et al 2018a). To select the optimal spectral variable subset, scholars have investigated varied methods such as gray correlation (GC) (Li et al 2016;Wang et al 2018b), stepwise regression (SR) (Zhang et al 2018) and variable importance in projection (VIP) (Qi et al 2017), and have achieved satisfactory effects. In addition, all the three methods have been widely applied in many studies, such as plant physiology, food engineering, mathematical statistics (Oussama et al, 2012;Maimaitiyiming et al 2017;Liu et al 2015).…”