Identifying Substitution items from transaction data
| Autor: | Ming-Jia Hsu, 許銘家 |
|---|---|
| Rok vydání: | 2018 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 106 Abstruct In the era of global economic highly competition, consumables of the world has gradually become the globalization. Consumers have to accept the new information fast all the time. Because of high prices and complicated products, consumers often have to face the choice of goods and substitution items. Consumers may face the problem of information overload. Therefore, enterprises not only can save the time that consumers gather product but improve the satisfaction of the consumer goods and to equip consumers' purchasing habits. They usually establish the commodity association rules. Association rules can be used for consumer preferences and identify possible consumer habits and the purchase of the portfolio and provide the right combination of goods to the consumer to look forward to increasing consumer willingness to buy. And the enterprise itself can effectively manage commodity and inventory costs. The traditional association of alternative rules mining is to find negative correlation for the main method, and the traditional association exploration of alternative rules always be considered resources and time-consuming. Therefore, the research to begin with to improve the computing performance and optimize the system as the main target. This visionary paper is only combination of traditional association rules and the formula of the economics of alternatives, then coupled with consumer behavior to find out the special of substitute goods and the possible combination of goods. Attempts to use the association rules from transaction database to analysis potential correlation between the combination of goods. And through a variety of association rules to find out the difference between items, and try to extract the portfolio of possible alternatives from the transaction data. In this paper, the mining out of the possible alternatives portfolio is be proved by the negative correlation in correlation coefficient. The study showed that utilize the concept of association rules together with the alternatives in economic and the consumer behavior, nearly 70% of the substitutes were subject to the negative correlation alternative rules; hence, it shows that this study has certain accuracy. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|