Application of Collaborative Filtering Analysis in Web Marketing
碩士 === 大同大學 === 資訊經營學系(所) === 107 === This paper describes an application of collaborative filtering recommendation analysis. Based on viewing behaviors, viewing time and periods, channels, film types, and the like of the viewers of an Internet TV, the HTV, collaborative search analysis algorithms a...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2019
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/4gxg4b |
| Summary: | 碩士 === 大同大學 === 資訊經營學系(所) === 107 === This paper describes an application of collaborative filtering recommendation analysis. Based on viewing behaviors, viewing time and periods, channels, film types, and the like of the viewers of an Internet TV, the HTV, collaborative search analysis algorithms are used to group the viewers and to export suitable film types and suitable contents to new users. Concerning the new customers, we obtain their related information when they registered on the website. The information helps us to recommend suitable contents so that the users will quickly find their favorite channels and increase adhesion. In addition, customer data are valuable to advertisers, creating an all-win situation. Our research focuses in particular on the cold start problem and the sparsity problem of the collaborative filtering algorithms. Through the frequency and queuing effects of online advertisement we may assess the impact of online TV on advertisement memory and purchase behaviors. Through collaborative filtering we may achieve better recommendations. Both user base and item base are used in the filtering to further analyze behavior patterns to achieve better recommendation precision.
|
|---|