Please use this identifier to cite or link to this item:
Title: 巨量資料技術於推薦系統之創新應用 --以Netflix 為例
Innovative Applications of Recommend System in Big Data - The Case of Netflix
Authors: 施人英
Jen-Ying Shih
Chu-Fen Liu
Keywords: 巨量資料
Big data
Recommend System
Issue Date: 2015
Abstract: 巨量資料目前已經普及到各種商業領域的運用,例如線上電子商務業者透過網路使用者的上網行為進行精準行銷、銷售;流通業用來作為消費者購物分析、電信業用以掌握行動用戶等。企業在巨量資料的應用目的,就是將資訊轉換為收益,在運用方式上,首先要瞭解企業面臨的議題,或者有一個新的概念想法要去驗證是否可行,之後針對相關議題進行資料收集。資訊儲存匯整後最重要的就是將這些資料進行處理分析,探索找尋出可能的趨勢或企業想要解答的問題。利基於既有的資料探勘技術結合不同分析演算法的推薦系統具有良好的發展方向和應用前景。推薦系統於瞭客戶的同時也提高了顧客對商務活動的滿意度,換來對商務網站的進一步支持使用。因此,近年來推薦系統在電子商務的應用越來越多,幾乎所有大型的電子商務及企業各線上服務系統,如Amazon、eBay、博客來、Netflix等,都有不同程度的使用了各種形式的推薦系統。各種提供個性化服務的網站也需要推薦系統的大力支持。在日趨激烈的競爭環境下,個人化推薦系統能有效挽留使用者,提高企業商業系統的銷售。成功的商業推薦系統會產生巨大的經濟效益,但隨著企業商業模式的進一步發展演進,不同企業在選用推薦系統也同樣面臨一系列挑戰。在選用推薦系統也同樣面臨一系列挑戰,本研究將以KDD(Knowledge Discovery and Data Mining)組織所研究發展演算理論及應用分析技術為基礎,探討如何透過不同推薦演算法技術如基於內容推薦法(Content-Based Recommendation)及協同過濾(Collaborative Filtering Recommendation) 挖掘巨量資料應用。其次,將以近年來所發展的出巨量資料應用技術,進行說明在商業服務的應用有效性與個案分析探討。文末將歸納上述探討結果,提出基於不斷演進的商務個人化服務技術的發展,社會與經濟價值終將創造更進一步的發展與提升。
The world has become excited about big data and advanced analytics not just because the data are big but also the potential impact driven by it. Many research reports have shown that the ability of mastermind the data analytics is one of the key foundations and driven factors that leads the successful companies such as Amazon, Google and Netflix (in which will be the case study in this research) ahead of others. Unfortunately for most of the conventional companies, thedata-analytics success merely refers to running a few tests such as focus group, calculation in Excel, or business model. Very few realize the real potential of the big data and advanced analytics. In this research, it shows that recommendation algorithms provide an effective form of targeted marketing by creating a personalized experience for each customer. For large retailers like, a good recommendation system is scalable over very large customer base and product catalogs. As one of the most successful approaches to building recommender systems, collaborative filtering uses the known preferences of a group of users to make recommendations or predictions of the unknown preferences for other users. It then presents three main categories of collaborative filtering techniques: memory-based, model-based, and hybrid collaborative filtering algorithms, with examples for representative algorithms of each category, and analysis of their predictive performance and their ability to address the challenges. The online video streaming company, Netflix, is one of the most reputed in utilizing collaborative filtering algorithms to predict user ratings for films based on previous ratings. This research has shown how Netflix has implemented a new generation of recommendation algorithms emerged, and demonstrates how the accuracy of prediction yields to the immeasurable influence of personalized recommendation hence benefiting its related business. Devices and ecosystem apps generate huge amounts of fast-moving data in a variety of forms nowadays and customers expect to receive value from the use of their data. A company that is skilled in analytics and can successfully convince its customers that it can provide such user value will out-win those who could not. Organizations that control and drive the most benefits from the data will eventually win.
Other Identifiers: G0102590141
Appears in Collections:學位論文

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.