摘 要生活观念的改变，促使着旅游业务领跑我国宏观经济，而随着计算机与互联网逐渐融入人们的新生活，更是让网上旅游占据着一枝独秀的地位。网上旅游产生的数据呈现指数型爆炸增长，其结构丰富多样，如何从海量数据中提取有用信息，根据用户特征和用户的喜好为用户推荐其感兴趣的旅游方式成为各个旅游网站的关键性挑战。所以，基于大数据的网上旅游推荐有着重要的研究意义。79778

本文研究的网上旅游推荐系统主要内容有：采用HADOOP生态框架，其高扩展性，高容错性，高效性的特点最为适合处理海量数据计算分析； 建立用户行为模型，根据用户的逐次输入的旅游信息，分析出用户的兴趣偏好，建立用户兴趣爱好模型，兴趣爱好模型数据将作为大权重数据加入推荐引擎算法；建立基于内容的加权型混合推荐算法，该算法组合了内容推荐和协同过滤推荐，除去了内容推荐算法依赖用户输入的旅游内容信息过强以及协同过滤算法推荐出的旅游线路多样性不足惊喜度低等弊端。同时该算法采用了回归模型，明显提高了对用户推荐旅游线路的推荐精度。

毕业论文关键词： 网上旅游线路；个性化推荐；HADOOP；混合推荐算法；HBASE

Abstract

Changes in the concept of life, led the tourism business to lead our macro economy, and with the computer and the Internet gradually into people's new life, but also to occupy a thriving online travel status。 The data generated by online tourism show exponential explosion growth, its structure is rich and varied, how to extract useful information from the massive data, according to the user characteristics and user preferences for users to recommend its interest in tourism has become a key challenge for various travel sites。 Therefore, based on large data online travel recommendation has important research significance。

In this paper, the main contents of the online tourism recommendation system contains: HADOOP ecological framework, its high scalability, high fault tolerance, high efficiency characteristics of the most suitable for dealing with massive data analysis and analysis; the establishment of user behavior model, according to the user's successively input tourism information , The user's interest preference is analyzed, the user interest model is established, and the interest model data will be added as the weighting data to the recommendation engine algorithm。 A content-based weighted hybrid recommendation algorithm is proposed。 The algorithm combines the content recommendation and collaborative filtering recommendation, The content recommendation algorithm depends on the user input of the tourism content information and collaborative filtering algorithm recommended the lack of persity of tourist routes and other shortcomings。 At the same time, the algorithm adopts the regression model, which obviously improves the recommendation accuracy of the recommended tourist routes。

Keywords: Online travel routes; Personalized recommendation; HADOOP; Hybrid recommendation algorithm; HBASE

目录

第1章 绪论 1。- 5 -

1。1 研究背景与意义 1。- 5 -

1。2 国内外研究水平 1。- 6 -

1。3 内容与结构 1。- 6 -

第2章 相关技术与分析 2。- 8 -

2。1 Hadoop生态圈平台架构分析