随着近年来网络信息化技术的迅猛发展，数据与信息的积累虽然为人们带来了诸多便利，但是由此带来的信息过载问题也是愈发严重。在旅游行业领域中，这类问题同样存在。近来个人自助游已渐成为主流趋势，用户希望根据自我喜好去规划个性化的旅游行程，而行程的定制工作往往需要人工过滤大量的相关信息，过程繁琐且工作量巨大。虽然目前市场上也有相关的行程定制服务，但是大多都采用人工对接，不仅费用高昂，且沟通成本较大。针对这些问题，本文设计了一款个性化自助游行程推荐系统。

在个性化自助游行程推荐工作中，合理的行程制定从用户角度下分析主要在于以下三个问题：首先是如何根据海量的用户评价信息保证行程中景点好坏的客观性；其次是如何根据用户特点的不同使行程更具个性化；最后就是如何使得行程的总体成本尽量低廉。

With the rapid development of network technology in recent years. Although the accumulation of data and information has brought a lot of benefits to people, the problem of information overload caused by it has become increasingly serious. Such problems also exist in the tourism industry. Recently, personal self-help travel has gradually become the mainstream trend. Users want to plan personalized travel itineraries according to their preferences. The itinerary customization often requires manual filtering of a large amount of relevant information, which is tedious and laborious. Although there are currently related itinerary customization services on the market, most of them use manual docking, which is not only expensive, but also has large communication costs. In response to these problems, this article has designed a personalized travel itinerary recommendation system.

In the personalized self-guided tour itinerary recommendation work, reasonable itinerary development from the perspective of the user mainly lies in the following three issues: The first is how to ensure the objectivity of the attractions in the itinerary based on the massive user evaluation information; the second is how to make the itinerary more personalized according to the different characteristics of the user; The last is how to make the overall cost of the trip as low as possible.

To solve these three main issues, this article introduces three technologies: Recommendation system, sentiment analysis and route planning. Then these three technologies are organically integrated into the system architecture composed of the Spark distributed computing engine and the Delta Lake storage framework. And carry out reasonable optimization and promotion. Thus, a personalized self-service travel itinerary recommendation system capable of supporting massive data analysis is completed.