An S-Curve Efficient Frontier For Evaluation Model

• Main Author: Salim, Fadzilah
• Format: Thesis
• Language: English; English
• Published: 2021
• Subjects: T Technology (General); TJ Mechanical engineering and machinery
• Online Access: http://eprints.utem.edu.my/id/eprint/25443/1/An%20S-Curve%20Efficient%20Frontier%20For%20Evaluation%20Model.pdf; http://eprints.utem.edu.my/id/eprint/25443/2/An%20S-Curve%20Efficient%20Frontier%20For%20Evaluation%20Model.pdf; http://eprints.utem.edu.my/id/eprint/25443/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=119917
• Institution: Universiti Teknikal Malaysia Melaka

A normal distribution has been an ideal model in providing mental picture on a distribution of an object. A normal distribution projects a global view on a variation of an object which is expected to spread. This mental picture signifies an ideal view on statistic motion in the past three centuries. A simple linear regression has been a practical predictive model following a normal distribution along an independent variable. While a linear model provides a compact support at a central mean, a distribution of a predicted value on both ends of a simple linear regression widens as an independent variable moves further away from its central mean. A simple linear regression gives a good fit on average for an expected predicted value. This present research study is looking at feasible maximum output along an independent variable called an efficient frontier. At the same time, a simple linear regression gives a more accurate prediction near a central mean. This research study focuses on a particular attention on both far ends of an efficient frontier curve. A non-linear model is presumed to perform better than a linear model. Upon overcoming the challenge of producing a practical non-linear model, a new S-curve efficient frontier for evaluation model is proposed in this study for better and practical estimation on an optimal output. An efficient frontier denotes an optimal curve in a nonlinear model. This new S-curve efficient frontier reflects a non-linear model from a single independent variable to a prediction valuation that has a positive first derivative throughout its progression. This non-linear S-curve model is proposed as an ideal projection in a scientific valuation model. Two types of quantitative secondary data have been collected, namely second-hand car prices and market equity share prices, to test on this model generation and validation. An S-curve efficient frontier model gives a better forecast along dynamic progress on an efficient frontier projection from a simple linear regression. Hence, this novel non-linear S-curve efficient frontier predictive for evaluation model may serve as an ideal projection to many real scenarios within positive derivative progression. More importantly, this S-curve model prescribes an ideal view on a statistical motion for future endeavours. An S-curve efficient frontier for evaluation model will provide and add a dynamic mental picture in addition to a normal distribution.