User profiling for personalized recommendation and user categorization

拟议的研究是在社交网络和数据分析领域，并应在加拿大ICT（信息和通信技术）部门的影响。EatSleepRIDE.com是一个面向摩托车骑手的在线社交网络平台。用户加入社区，发布/分享他们的故事，新闻，照片，评论等，并阅读以及评论其他人的帖子。该项目旨在解决两个具体问题。一是信息过载 问题.只有热门和最新的帖子出现在首页。用户很难找到他们可能喜欢的其他帖子。ESR希望找到一种解决方案，向用户推荐帖子。第二个问题是如何从ESR社区为第三方生成可操作的和有价值的见解，以促进针对个人和用户群体的有针对性的广告或营销。 在这个项目中，我们建议应用用户建模技术来构建丰富的用户配置文件，其中包含来自多个来源的信息，包括用户在注册时提供的信息，用户的社交网络活动（帖子，喜欢，评论，关注），用户的骑行历史等。使用生成的用户配置文件，我们可以构建个性化推荐系统，为ESR用户提供相关和高质量的信息，其可以解决上述第一个问题。为了解决第二个问题，我们可以根据用户的个人资料将其分类为预定义的组（例如新用户与有经验的用户）或自组织的组（即组中的类似用户），这可以帮助ESR从更有针对性的广告中获得收入。拟议的项目将通过将乘客与旅游目的地相匹配，使加拿大的旅游业受益。它还将通过提供一个更加个性化的社交平台，使社区成员之间更容易分享信息，并改善骑手的在线和潜在骑行体验，使摩托车手社区受益。它可以帮助ESR和其他第三方开展更有针对性的营销活动，这将为相关的加拿大公司和旅游运营商带来经济利益。

The proposed research is in the area of social network and data analytics, and should have an impact in Canadian ICT (Information and Communications Technology) sector. EatSleepRIDE.com (ESR) is an online social networking platform for motorcycle riders. Users join the community, post/share their stories, news, photos, reviews, etc. and read as well as give comments to others' posts. There are two specific problems this project is intended to solve. One is the information overload problem. Only the hot and latest posts appear in the front page. It is not very easy for users to find other posts they might like. ESR wants to find a solution to recommend the posts to their users. A second problem is how to generate actionable and valuable insights from the ESR community for third parties to facilitate targeted advertising or the marketing for individuals as well as user groups. In this project, we propose to apply the user modeling technique to build enriched user profiles with information from multiple sources, including the user-provided information during registration, users' social network activities (posts, likes, comments, follows), users' riding histories, etc. Using the generated user profiles, we could build the personalized recommender system to provide the relevant and quality information to ESR users, which can solve the above-mentioned first problem. To solve the second problem, we could categorize users based on their profiles into pre-defined groups (e.g. new vs. experienced riders) or self-organized groups (i.e. similar users in a group), which could help generate revenue for ESR from more targeted advertisements. The proposed project will benefit Canada's tourism sector by matching riders with tourism destination. It will also benefit the motor-rider community by offering a more personalized social platform for easier information sharing between community members, and improved online and potentially riding experiences for riders. It could help ESR and other third parties launch more targeted marketing campaigns, which will bring economic benefits to related Canadian companies and tourism operators.