碩士 / 國立中山大學 / 金融創新產業碩士專班 / 106 / International trade is based on shipping industry. Transportation of goods all around the world is contingent on this industry, so it’s affect by demand and supply of global trade. If the volume of trading reduced, it might imply that the economic is during a regression period and reflect directly to the revenue of the industry. As a result, this thesis focus on predicting the stock price of shipping industry. At the meanwhile, we intent to find out the macroeconomic factors which have a relative important effect on the stock price.
This thesis transformed macroeconomic factors to trend period and set rolling window to capture the momentum of long-term trend of economic. Then we conduct feature selection to find out critical factors and forecast the stock price through the model building of LASSO regression.
We used back-testing to examine the effectivity of the result. It shows the win-rate is up to 69.92% and the annual return rate is 12.08% with basic model. The win-rate is up to 78.02% and the annual return rate is 12.21% with threshold adjustment of 0.01, better than holding the shipping index continuously.
The following is conclusion of this thesis: Firstly, macroeconomic factors do have the ability of forecasting the stock price of shipping index. Second, the major factors that affect the shipping industry are annual growth rate of M2 in Taiwan, initial application for unemployment benefits in US, Taiwan Capitalization Weighted Stock Index, annual growth rate of M1 in China, CRB Futures Price Index, Index of export in US, and ISM Production index in US. Third, shipping industry in Taiwan is a relative efficiency market and often changes its construction.
Keywords: shipping industry, forecasting stock price, macroeconomic factors, LASSO regression
| Identifer | oai:union.ndltd.org:TW/106NSYS5457053 |
| Date | January 2018 |
| Creators | Chia-Wei Hung, 洪嘉蔚 |
| Contributors | Chang-Chiang Chin, 秦長強 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 42 |
Page generated in 0.017 seconds