Papers by Keyword: Partial Least Square Regression (PLSR)

Abstract: Heavy metals pollution in agricultural soils has been an important problem to human health, mapping large-scale spatial distribution of soil heavy metals is urgently needed. Instead of traditional methods, time-consuming and destructive, soil properties predicted by remote sensing technology shows a lot of advantages, which makes large area of real-time dynamic monitoring as possible. However, before achieving prediction using spectra data, the first thing to do is that finding the spectral characteristics of soil heavy metals. In this paper, taking Cr and Cu for example, the correlations between soil heavy metals content and laboratory-measured reflectance is studied using partial least squares regression (PLSR), which is an adaptive method to examine linear between spectrum and concentration. First of all, using the raw spectra, remove outliers of heavy metals concentration by PLSR modeling. Next, though comparing RMSEC and RMSEV against PLSR components, and cumulative explanatory of spectral components to metal content using different pre-precessing methods, find the right pre-pcocessing is CR and optimum number of components to Cr and Cu are 3 and 2 respectively. Simultaneously, with the meaning of PLSR models regression coefficients, we analysis the spectral characteristics of Cr and Cu, although can not to realize the prediction only take use of these spectra, which is still essential to achieve simulating spatial distribution of soil heavy metal by remote sensing.

Abstract: Based on partial least-squares regression taking into account interactional items among independent variables, this paper had a prediction on concrete strength at the 28th day. Taking proportion of flyash in cementing material, usage amount of cementing material, ash-water ratio as independent variables , and concrete strength at the 28th day as dependent variable , the forecast model of concrete strength was obtained. It was found that press residual value decreased with the increase of number of latent variables, and number of latent variables were three according to Press residual value versus number of latent variables. The normal regression coefficient of ash-water ratio was the largest in three influence factors, this indicated that the influence of ash-water ratio was largest to concrete strength at the 28th day; The determination coefficient of forecast model obtained in this paper was 0.9353, the error of forecast model was. The following conclusion can be drawn that, the model is accurate and credible, and the partial least-squares regression taking into account interactional items among independent variables is a eximious non-linear method, and it is worthy to spread its application in the forecast analysis of concrete strength at the 28th day.

Abstract: In situ determination of optimal harvest time of tomatoes is of value for growers to optimize fruit picking schedule. This study evaluates the feasibility of using visible and near infrared (VIS-NIR) spectroscopy to make an intact estimation of harvest time of tomatoes. A mobile, fibre-type, AgroSpec VIS-NIR spectrophotometer (Tec5, Germany), with a spectral range of 350-2200 nm, was used for spectral acquisition of tomatoes in reflection mode. The harvest time of tomatoes was measured by the days before harvest. After dividing spectra into a calibration set (70%) and an independent prediction set (30%), spectra in the calibration set were subjected to a partial least-squares regression (PLSR) with leave-one-out cross validation to establish calibration models. Validation of calibration models on the independent prediction set indicates that the best model can produce excellent prediction accuracy with coefficient of determination (R2) of 0.90, root-mean-square error of prediction (RMSEP) of 2.5 days and residual prediction deviation (RPD) of 3.01. It is concluded that VIS-NIR spectroscopy coupled with PLSR models can be adopted successfully for in situ determination of optimal harvest time of tomatoes, which allows for automatic fruit harvest by a horticultural robot.

Abstract: This study used visible and near-infrared (VIS-NIR) spectroscopy for size estimation of tomato fruits of three cultivars. A mobile, fibre-type, VIS-NIR spectrophotometer (AgroSpec, Tec 5, Germany) with spectral range of 350-2200 nm, was used to measure reflectance spectra of on-vine tomatoes growing from July to September 2010. Spectra were divided into a calibration set (75%) and an independent validation set (25%). A partial least squares regression (PLSR) with leave-one-out cross validation was adopted to establish calibration models between fruit diameter and spectra. Furthermore, the latent variables (LVs) obtained from PLS regression was used as input to back-propagation artificial neural network (BPANN) analysis. Result shows that the prediction of PLSR model can produce good performance with coefficient of determination (R²) of 0.82, root-mean-square error of prediction (RMSEP) of 4.87 mm and residual prediction deviation (RPD) of 2.35. Compared to the PLSR model, the PLS-BPANN model provides considerably higher prediction performance with R² of 0.88, RMSEP of 3.98 mm and RPD of 2.89. It is concluded that VIS-NIR spectroscopy coupled with PLS-BPANN can be adopted successfully for size estimation of tomato fruits.

Abstract: In this research, the wood residues of mixed birch and larix gmelini rupr are used as the bulk material, and phenolic resin glue and moso bamboo slice as reinforcing material to make flakeboard through three-phase pressure process. Data Processing System (DPS) is used to analyze glue volume, density and other technical parameters’ impact on the performance of wood composite product, establish multiple linear regressive mathematic models, compare and optimize experiment plan, find out and verify the best technical conditions. The result shows that the overall objective function will reach 6.908 when glue volume is 8.107%, density 0.870 g/cm3, thickness 2.015cm and moisture content of 9.799%.

Abstract: Mechanical parameter analysis is a complicated issue since it is influenced by many factors. Closely related with the influencing factors of compressibility coefficients of rock material (sandstone), this article first introduces the way to process partial least square regression (PLSR) analysis. The process of carrying out PLSR is divided into six steps as for analysis and prediction of the regression model, which are data preparation, principle collection, regression model for first principle component, secondary principle analysis, establishment of final regression model and number determination of principal component l. And then introduces PLSR for application of analysis and prediction of compressibility coefficients with 30 experiment samples. Seven prediction samples are carried out by PLSR with the training process of 30 samples. The result shows PLSR has good accuracy in prediction under the condition that the model is properly deprived based on certain experimental samples. Finally, some conclusions are made for further study on both mechanical parameters and partial least square regression method.

Abstract: Based on remote sensing and field measured data, this paper discussed the effectiveness of selecting optimal variables with a cross-variable by VIP≥1 and used partial least squares regression (PLSR) to build a model for estimating surface soil organic carbon of burned area. The results showed that: variables such as K-T1, K-T2, TM2, TM4, and TM5/TM4 have larger contribution to the model, their VIP values were 1.5116, 1.1915, 1.3545, 1.2242 and 1.4275, respectively. The intensity index of Tasseled Cap transformation has a higher value than the original bands of TM2 and TM4, and the contribution of bands ratio is higher than the single band. Further, the PLS model was applied to estimating the special distribution of soil organic carbon (SOC) content (unit in g.kg^-1) in entire study area. We conclude that different fire intensity affected on soil organic carbon variously, and followed an order as high intensity>moderate intensity>light intensity. The average density of surface soil organic carbon was 12.53 kg per m², soil organic carbon storage was 202.46×10⁶kg, the fire-released soil organic carbon was 29.8×10⁶kg in study area.

Abstract: Algae bloom in the Tai Lake is a major issue and affects the water supply to the surrounding cities greatly. Chlorophyll a (Chl-a) is a common indicator that represent the trophic status in lakes. Spatial and temporal variations of Chl-a concentration are analyzed on the basis of sample data at 21 sites during the period of 2001 to 2005. Data at the sites located in the Meiliang Bay, Zhushan and Wulihu show greater fluctuations than that at other sites. A general trend showing that high values in northern part and low values in southern part of the Tai Lake is observed in seasonal mean values of Chl-a concentration for four seasons. Most high Chl-a concentrations occurred in summer (June to August) and autumn (September to November). Quantitative relationships between Chl-a and other water quality factors are investigated at all sites. Relative good relationships are obtained between Chl-a concentration and other water quality factors during 2001 to 2004 by using partial least squared regression. Prediction of Chl-a concentration in 2005 with above models produce worse results, which may be due to the occurrence of some extreme high values of Chl-a concentration in that year. Even though, acceptable predictions are obtained at several sites. Since the water quality in the lake is affected greatly by the inflow of nutrients from rivers, these relationships will be helpful for monitoring Chl-a variation with the combination of hydrological models that is able to simulate the inflow of nutrients.

Abstract: Fast measurement of soil properties is of great value for farmland fertilizer management with the purpose of reducing environmental pollution caused by over-fertilizing and improving yield/cost ratio of land. In the study, an optical fiber sensing system was introduced. It consisted of a topsoil opener, a portable spectrophotometer (350-2500nm) and an optical fiber. Prediction models of total nitrogen (N), phosphorous (P), moisture content (MC), pH and organic matter (OM) were created in laboratory and validated by field testing. Soil samples were collected from 30 farmland blocks covering north of Zhejiang Province and south of Jiangsu Province, P.R. China. Partial least squares (PLS) regression with full cross-validation was used for calibration models building. The correlation coefficient of R2 between on-line and laboratory measurements of N, P, MC, pH and OM are 0.92, 0.78, 0.85, 0.83 and 0.76, respectively. The study shows the potential of using the optical fiber soil sensing system for next generation of precision farming.

Abstract: In this paper, a kind of SMA model based on hysteretic nonlinear theory was developed. Von del Pol nonlinear difference item was introduced to interpret the hysteresis phenomenon of strain-stress curve of SMA. The coupling relationship between strain and temperature was obtained in partial least-square regression method to describe the variation of stiffness with temperature. Based on above, the final relationship among strain, stress and temperature was set up. The result of significance test shows that the final model can describe the characteristics of SMA in different temperature well. The new SMA model broadens the region of temperature, and is easy to be analyzed in theory, which is helpful to application of SMA in engineering fields.