Significant factors influencing the rate of cigarettes burn

Chew, Peng (2015) Significant factors influencing the rate of cigarettes burn. Universiti Malaysia Sabah. (Unpublished)

Significant factors influencing the rate.pdf

Download (10MB) | Preview


Background: Tobacco is the green leaf that was grown in warm climates. However, cigarette is the most common method of consuming tobacco and contains many different chemicals. Cigarette smoking was such a strong addition that was very difficult to break. Therefore, a study is constructed to determine which chemical constituents in tobacco leaf had a most significant influence on the rate of cigarettes burn. Main Methods: The multiple regression analysis was introduced in this study, because the dependent variable was single quantitative dependent variable. Preliminary, the data was filtered by running through the factor analysis and data transformation. The filtered data then was divided into three partitions: model building approach, verification by using MAPE and estimating the missing values. To obtain the best model, in estimating the 4 phases would be carried out in model building procedure, where NPC test, multicollinearity test based on VIF test and coefficient test would be carried out in Phase2. Main Findings: Eight selection criteria test had been explored and introduced in the process of getting the best model, where the best model is found when it satisfied all of the minimum values of the eight selection criteria. From the best model, there were 3 variables emerged which had been included one single independent variable of calcium and two first-order interaction variables such as, calcium interact with potassium and chlorine interact with potassium. Moreover, based on the result in second partition, P2, the MAPE value for the best model is 7.73%. Conclusion: Calcium, potassium and chlorine were found to be the most significant factors in estimating the rate of cigarettes burn. While, the estimation efficiency (MAPE) results tend to signify that model is accepted.

Item Type: Academic Exercise
Uncontrolled Keywords: multiple regression, model building procedure,multicollinearity, VIF, coefficient test, best model, estimation efficiency
Subjects: Q Science > QA Mathematics
S Agriculture > SB Plant culture
Divisions: FACULTY > Faculty of Science and Natural Resources
Depositing User: Munira
Date Deposited: 03 Feb 2018 21:52
Last Modified: 03 Feb 2018 21:52

Actions (login required)

View Item View Item