A Bayesian Forecasting Model for Market Share – Take the Database of IT Industry as an Example

• Main Authors: Che-Yu Hung, 洪哲瑜
• Other Authors: Li-Chung Jen
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/mhbb7z

碩士 === 國立臺灣大學 === 國際企業學研究所 === 106 === With technology advanced, commercial war on channel spreads not only physically but also virtually while customers get the sweet. Customers can easily more and more product so as to raise their standards to another level. How can a leader manage to gain profit in an environment filled with vicious competitor and fickle customer demands. In order to achieve this goal, manager should efficiently allocate resource. In addition to knowing their own products and advantages, manager should keep an eye on other competitor and the competitive environment. However, the gathering of the information is difficult task. This often lead to bad decisions from managers due to wrong numbers or data. In order to solve this problem, this research aims to develop a highly customized Bayesian Dynamic Model to track sale numbers in real-time and accurate basis. This model lets managers combine internal complete information and external incomplete competitor data and provide an instant review into market and competitive environment to help better the decision quality. This research adopt the data base from W Company, and take PC sales data from 2014 to 2016 as research number. This research takes computer suppliers, sale regions and sale channels to discuss interconnection between the variables, and create a relation matrix to predict market share in next season.