Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Development of Regression Models for the Prediction of Leaf Fresh Weight of Aloe Vera

Md. Mahabubul Alam
Department of Crop Botany Bangladesh Agricultural University Mymensingh-2202, Bangladesh.

Dr. MSA Fakir
Professor
Department of Crop Botany, Bangladesh Agricultural University Mymensingh-2202, Bangladesh.

Dr. Md. Rashedur Rahman
Associate Professor
Department of Agronomy, Bangladesh Agricultural University, ymensingh-2202, Bangladesh.

Dr. Md. Asaduzzaman
Scientific Officer
Olericulture Division, Horticulture Research Center,Bangladesh Agricultural Research Institute. Joydebpur, Gazipur-1701, BANGLADESH

Abstract/ Executive Summary:

An experiment was conducted in Nilphamari district, Bangladesh, during September to December, 2004 to develop and test linear regression models for predicting leaf fresh weight of Aloe vera from simple measurements. The experiment was designed as RCBD with three replications. There were three row to row spacing (40, 30 and 20 cm) and three plant to plant spacing (25, 40 and 60 cm), therefore, the total number of plots was 27. The size of the unit plot was 4 m × 2.5 m. After transplanting of the sucker, all recommended intercultural operations were done till harvest. Two hundred forty three plants from three destructive harvests at different ages were collected. In each harvest, 81 plants were sampled. Weight, length, breadth, thickness and area of each leaf were measured. Fifteen regression models were derived for predicting leaf fresh weight (LFwt) in Aloe vera. High R², low MSE and small percentage of deviation are considered for selection indices in regression models of estimation of any LFwt measurement. Four models, LFwt = 2.7814 LL – 30.94 0, LFwt = 7.3118LB² + 0.0618, LFwt = 111.7400LT – 53.622 and LFwt = 1.0297LA – 3.544 showed higher R² value, lower MSE and minimum percentages of deviation (0.8619, 14.8611 & 0.140; 0.9448, 6.9940 & 0.004; 0.8246, 18.7990 & 0.001 and 0.9553, 6.1426 & 0.026, respectively). These models were tested for accuracy by plotting actual leaf fresh weight (LFwt) versus estimated (predicted) leaf fresh weight (LFwt) using regression models and result showed that there was a high co-linearity. Therefore, these models could be used to estimate leaf fresh weight (LFwt).

Keywords: Regression model, Leaf fresh weight, Aloe vera

Location: Experimental Field Laboratory, Department of Crop Botany, BAU, Mymensingh, Bangladesh

Start Date: 01-09-2004

End Date: 31-12-2004

Program Area: Development of Host and Medicinal Plants

Commodity: Aloe Vera

Objectives:

The study was carried out to develop and test regression models that would enable us in predicting leaf fresh weight from simple linear measurements (length, breadth, thickness, leaf area etc.) of leaf.

Methodology:

Planting material and cultivation procedures: Four months old suckers of Aloe vera plants were used as planting material. The suckers were collected from Haybotpur-Laxipur of Natore Sadar district of Bangladesh. The suckers were transplanted in the well prepared and fertilized land. The dead suckers were replaced with new suckers. Initially the plots were irrigated and plants were clipped off for old leaves leaving only 4 to 5 leaves.

Crop sampling and data collection: Three destructive harvesting were made at different stages and thus a total of 243 plants were obtained. In the first harvest, 81 randomly selected plants, three from each of 27 plots, were collected at 1 month after planting. Similarly the second and third harvesting were done at 2 and 3 months after planting.

Measurement of leaf size, leaf area and leaf weight: The leaves were collected from the sampled plant and leaf length (LL), leaf breadth (LB) and leaf thickness (LT) were measured with the help of a cm scale for measuring leaf area manually (Fig. 1). Leaf area (LA) was measured in cm² by multiplying leaf base breadth with leaf length and ½. The fresh weight of leaves were obtained trough a digital balance. The leaves were dried in the oven at 80 ± 2⁰C until constant weight (48 hours) was result.

Calculation and prediction of leaf fresh weight (LFwt): Regression model of predicting leaf fresh weight (LFwt) from leaf length (LL) was developed (LFwt = a + bLL). Actual LFwt by digital balance was also recorded against each estimated LFwt. Further, regression of actual LFwt versus estimated LFwt was performed. Thus, predication of regression equation of LFwt from leaf length (LL) was assessed. Regression models for predicting LFwt from leaf breadth (LB) and leaf thickness (LT) were also developed and was tested (LFwt = a + bLB) like leaf length (LL). Actual leaf area of a particular leaf was correlated with leaf fresh weight (LFwt) and thus regression equation (LFwt = a + bLA) for the prediction of LFwt from LA was developed. Actual leaf fresh weight was recorded by a digital balance against each estimated LFwt. Further, regression analysis of actual LFwt versus estimated LFwt was performed. Thus, an equation for the prediction of leaf fresh weight from the leaf area was tested. Regression models for predicting LFwt from leaf dry weight (LDwt) were developed and was tested (LFwt = a + bLDwt) like leaf area (LA).

Source of Information: Intl. J. BioRes. 1(1): 35-42, July, 2006

Article Link (Online Journal):

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

Fifteen regression models were derived for predicting leaf fresh weight (LFwt) in Aloe vera. High R², low MSE and small percentage of deviation are considered for selection indices in regression models of estimation of any LFwt measurement. Four models, LFwt = 2.7814 LL – 30.94 0, LFwt = 7.3118LB² + 0.0618, LFwt = 111.7400LT – 53.622 and LFwt = 1.0297LA – 3.544 showed higher R² value, lower MSE and minimum percentages of deviation (0.8619, 14.8611 & 0.140; 0.9448, 6.9940 & 0.004; 0.8246, 18.7990 & 0.001 and 0.9553, 6.1426 & 0.026, respectively). These models were tested for accuracy by plotting actual leaf fresh weight (LFwt) versus estimated (predicted) leaf fresh weight (LFwt) using regression models and result showed that there was a high co-linearity. Therefore, these models could be used to estimate leaf fresh weight (LFwt).

Knowledge Management: Journal