Modelling procedures in determining heavy metals concentration: A case study using barks of the cinnamon tree
High concentrations of heavy metals may present as toxins to living organisms. Hence, heavy metal absorption by the cinnamon tree (Cinnamomum iners) grown in Universiti Malaysia Sabah (UMS), Malaysia was investigated in order to assess its composition, concentration and dynamics. The relationship of...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Researchgate
2016
|
| Subjects: | |
| Online Access: | https://eprints.ums.edu.my/id/eprint/44473/1/FULL%20TEXT.pdf https://eprints.ums.edu.my/id/eprint/44473/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | High concentrations of heavy metals may present as toxins to living organisms. Hence, heavy metal absorption by the cinnamon tree (Cinnamomum iners) grown in Universiti Malaysia Sabah (UMS), Malaysia was investigated in order to assess its composition, concentration and dynamics. The relationship of heavy metals concentration in barks of C.iners was determined using the multiple regression (MR) technique. The model building procedures were illustrated and discussed. Five independent variables were considered during field and experimental data collection which were namely, diameter of breast height, stem height, average ppm in bark, average ppm in soil and concentration of heavy metal in soil. The concentrations of heavy metals in the bark form the six dependent variables, and they were Cadmium (Cd), Copper (Cu), Iron (Fe), Lead (Pb), Nickel (Ni) and Zinc (Zn). The non-parametric bootstrapping method was used to generate the small sample size (n=28) into 500 observations. The 80 multiple regression (MR) models were developed up to the fourth-order interactions. Results obtained were subjected to statistical modelling, enhanced by the four phase model-building procedures and the process of getting the best model based on the eight selection criteria (8SC). The progressive elimination of variables using the variance inflation factor (VIF) was used to remove collinearity variables from these models. The forecasting criteria of mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE) were compared and discussed. Comparisons were made to have the best model equation from the six respective heavy metals concentrations. The model M73.20.0 with the toxic heavy metal concentration of Copper (Cu) was obtained as the best model. This thus indicated that Copper was the most toxic heavy metal concentration absorbed in this bark of C. iners. |
|---|