As social media sites grow in popularity, tagging has naturally emerged as a method of searching, categorizing and filtering online information, especially multimedia content. The unrestricted vocabulary users choose from to annotate content however, has often lead to an explosion of the size of space in which search is performed. This paper is concerned with investigating generative models of social annotations, and testing their efficiency with respect to two information consumption oriented tasks. One task considers recommending new tags (similarly new resources) for new, previously unknown users. We use perplexity as a standard measure for estimating the generalization performance of a probabilistic model. The second task is aimed at recommending new users to connect with. In this task, we examine which users' activity is most discriminative in predicting social ties: annotation (i.e. tags), resource usage (i.e. artists), or collective annotation of resources altogether. For the second task, we propose a framework to integrate the modeling of social annotations with network proximity. The proposed approach consists of two steps: (1) discovering salient topics that characterize users, resources and annotations; and (2) enhancing the recommendation power of such models by incorporating social clues from the immediate neighborhood of users. In particular, we propose four classification schemes for social link recommendation, which we evaluate on a real--world dataset from Last.fm. Our results demonstrate significant improvements over traditional approaches.
{"title":"Exploring generative models of tripartite graphs for recommendation in social media","authors":"C. Chelmis, V. Prasanna","doi":"10.1145/2463656.2463658","DOIUrl":"https://doi.org/10.1145/2463656.2463658","url":null,"abstract":"As social media sites grow in popularity, tagging has naturally emerged as a method of searching, categorizing and filtering online information, especially multimedia content. The unrestricted vocabulary users choose from to annotate content however, has often lead to an explosion of the size of space in which search is performed. This paper is concerned with investigating generative models of social annotations, and testing their efficiency with respect to two information consumption oriented tasks. One task considers recommending new tags (similarly new resources) for new, previously unknown users. We use perplexity as a standard measure for estimating the generalization performance of a probabilistic model. The second task is aimed at recommending new users to connect with. In this task, we examine which users' activity is most discriminative in predicting social ties: annotation (i.e. tags), resource usage (i.e. artists), or collective annotation of resources altogether. For the second task, we propose a framework to integrate the modeling of social annotations with network proximity. The proposed approach consists of two steps: (1) discovering salient topics that characterize users, resources and annotations; and (2) enhancing the recommendation power of such models by incorporating social clues from the immediate neighborhood of users. In particular, we propose four classification schemes for social link recommendation, which we evaluate on a real--world dataset from Last.fm. Our results demonstrate significant improvements over traditional approaches.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126133841","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Although considerable amount of work has been conducted recently of how to predict links between users in online social media, studies exploiting different kinds of knowledge sources for the link prediction problem are rare. In this paper latest results of a project are presented that studies the extent to which interactions -- in our case directed and bi-directed message communication -- between users in online social networks can be predicted by looking at features obtained from social network and position data. To that end, we conducted two experiments in the virtual world of Second Life. As our results reveal, position data features are a great source to predict interacts between users in online social networks and outperform social network features significantly. However, if we try to predict reciprocal message communication between users, social network features seem to be superior.
{"title":"Predicting interactions in online social networks: an experiment in Second Life","authors":"Michael Steurer, C. Trattner","doi":"10.1145/2463656.2463661","DOIUrl":"https://doi.org/10.1145/2463656.2463661","url":null,"abstract":"Although considerable amount of work has been conducted recently of how to predict links between users in online social media, studies exploiting different kinds of knowledge sources for the link prediction problem are rare. In this paper latest results of a project are presented that studies the extent to which interactions -- in our case directed and bi-directed message communication -- between users in online social networks can be predicted by looking at features obtained from social network and position data. To that end, we conducted two experiments in the virtual world of Second Life. As our results reveal, position data features are a great source to predict interacts between users in online social networks and outperform social network features significantly. However, if we try to predict reciprocal message communication between users, social network features seem to be superior.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127588021","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. Heupel, S. Scerri, Mohamed Bourimi, D. Kesdogan
Recommender systems depend on the amount of available and processable information for a given purpose. Trends towards decentralized online social networks (OSNs), promising more user control by means of privacy preserving mechanisms, lead to new challenges for (social) recommender systems. Information, recommender algorithms rely on, is no longer available, (i.e. central user registries, friends of friends), thus shared data is reduced and centralized processing becomes difficult. In this paper we address such drawbacks based on identified needs in the decentralized OSN di.me and present concepts overcoming those for selected functionalities. Besides this, we tackle the support of privacy advisory, warning the user of risks when sharing data.
{"title":"Privacy-preserving concepts for supporting recommendations in decentralized OSNs","authors":"M. Heupel, S. Scerri, Mohamed Bourimi, D. Kesdogan","doi":"10.1145/2463656.2463659","DOIUrl":"https://doi.org/10.1145/2463656.2463659","url":null,"abstract":"Recommender systems depend on the amount of available and processable information for a given purpose. Trends towards decentralized online social networks (OSNs), promising more user control by means of privacy preserving mechanisms, lead to new challenges for (social) recommender systems. Information, recommender algorithms rely on, is no longer available, (i.e. central user registries, friends of friends), thus shared data is reduced and centralized processing becomes difficult. In this paper we address such drawbacks based on identified needs in the decentralized OSN di.me and present concepts overcoming those for selected functionalities. Besides this, we tackle the support of privacy advisory, warning the user of risks when sharing data.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123179449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Social media as well as mobile devices have woven themselves into everyday life, mediating various implicit and explicit social interactions. The analysis and modeling of the interaction data, including both physical and online social interactions is receiving increasing interest. A prerequisite is then given by effective approaches for data collection, covering both sensor data and social media artifacts. This paper describes the Sensor Data Collection Framework (SDCF), a compact, versatile and easily extensible open source framework for mobile sensing and ubiquitous data collection. It provides an overview on core concepts and architecture. Furthermore, we discuss first experiences and results of applying the framework in a collaborative workgroup context.
{"title":"Towards capturing social interactions with SDCF: an extensible framework for mobile sensing and ubiquitous data collection","authors":"M. Atzmüller, Katy Hilgenberg","doi":"10.1145/2463656.2463662","DOIUrl":"https://doi.org/10.1145/2463656.2463662","url":null,"abstract":"Social media as well as mobile devices have woven themselves into everyday life, mediating various implicit and explicit social interactions. The analysis and modeling of the interaction data, including both physical and online social interactions is receiving increasing interest. A prerequisite is then given by effective approaches for data collection, covering both sensor data and social media artifacts. This paper describes the Sensor Data Collection Framework (SDCF), a compact, versatile and easily extensible open source framework for mobile sensing and ubiquitous data collection. It provides an overview on core concepts and architecture. Furthermore, we discuss first experiences and results of applying the framework in a collaborative workgroup context.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"199 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131813221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We receive many recommendations of friends in online social networks such as Facebook and LinkedIn. These friend recommendations are based usually on common friends or similar profile such as having the same interest or coming from the same company, a trait known as homophily. However, many times people do not know why they should add this friend. Should I add this friend because we met from a conference and if so, what conference? Existing friend recommendation systems cannot answer this question easily. In this paper, we create a friend recommendation system using proximity and homophily, that we conduct in the workplace and conference. Besides common friends and common interests (homophily features), we also include encounters and meetings (proximity features) and messages sent and question and answer posts (social interaction features) as reasons for adding this person as a friend. We conduct a user study to examine whether our friend recommendation is better than common friends. Results show that on average, our algorithm recommends more friends to participants that they add and more recommendations are ranked as good, compared with the common friend algorithm. In addition, people add friends due to having encountered them before in real life. The results can be used to help design context-aware recommendations in physical environments and in online social networks.
{"title":"Who should I add as a \"friend\"?: a study of friend recommendations using proximity and homophily","authors":"Alvin Chin, Bin Xu, Hao Wang","doi":"10.1145/2463656.2463663","DOIUrl":"https://doi.org/10.1145/2463656.2463663","url":null,"abstract":"We receive many recommendations of friends in online social networks such as Facebook and LinkedIn. These friend recommendations are based usually on common friends or similar profile such as having the same interest or coming from the same company, a trait known as homophily. However, many times people do not know why they should add this friend. Should I add this friend because we met from a conference and if so, what conference? Existing friend recommendation systems cannot answer this question easily. In this paper, we create a friend recommendation system using proximity and homophily, that we conduct in the workplace and conference. Besides common friends and common interests (homophily features), we also include encounters and meetings (proximity features) and messages sent and question and answer posts (social interaction features) as reasons for adding this person as a friend. We conduct a user study to examine whether our friend recommendation is better than common friends. Results show that on average, our algorithm recommends more friends to participants that they add and more recommendations are ranked as good, compared with the common friend algorithm. In addition, people add friends due to having encountered them before in real life. The results can be used to help design context-aware recommendations in physical environments and in online social networks.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"57 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120903344","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Social networks have evolved over the last decade into an omni-popular phenomenon that revolutionized both the online and offline interactions. They are used for a variety of purposes and are fast becoming the place to share and discover news, activities, and content of interest. Facebook alone reports more than 1 billion users, each having on average 130 friends and connected to 80 communities, and spending on Facebook less than one hour a day. The volume of the generated content of potential interest is, thus, overwhelming and ever growing, but the time spent on the social networks is fairly limited. How can users stay abreast of the activities of interest given this severe information overload? Activity feed is a simple mechanism deployed nowadays by many social networks, which performs information filtering on the users' behalf. Typically, activity feed encompasses reverse chronologically ordered items corresponding to activities carried out by direct friends and followees. However, activity feed can hardly cope with the volume and diversity of the activities. In order to alleviate information overload, simplify content discovery, and sustain user engagement, there is a need to personalise the activity feed, i.e., identify items of a particular interest and relevance for the user and filter out irrelevant items. The feed personalisation task can be naturally represented as a top-K recommendation problem. Let us denote by N the set of items corresponding to activities that can potentially be included in the feed, e.g., all the activities carried out since the user's last visit. Hence, the personalisation task aims at selecting and recommending a smaller set of items, K (|K|≪|N|), corresponding to activities of the highest relevance for the user. Essentially, the recommendation process entails scoring all the |N| candidate items and selecting |K| top-scoring items. What information can facilitate the item scoring? When interacting with a social network, users typically leave very little explicit feedback, primarily their 'likes'. There is a moderate amount of strong implicit user-to-user feedback, e.g., friending and direct communication (messages and comments), and abundance of weak implicit user-to-activity feedback, such as content viewing and contribution, community membership, and event participation. Finally, there is some self-reported and often unreliable information pertaining to user demographics, location, preferences, skills, or interests. How can all this this information be modelled, fused, mined, and eventually leveraged for scoring and recommending activity feed items? This problem has been investigated from different angles in the recent years [1-10]. In this talk, we will overview most prominent works into the personalisation of the activity feed. These works proposed a spectrum of algorithmic approaches and evaluated them with numerous social networks of a highly heterogeneous nature. We will summarise the main components
{"title":"Network activity feed: finding needles in a haystack","authors":"S. Berkovsky","doi":"10.1145/2463656.2463657","DOIUrl":"https://doi.org/10.1145/2463656.2463657","url":null,"abstract":"Social networks have evolved over the last decade into an omni-popular phenomenon that revolutionized both the online and offline interactions. They are used for a variety of purposes and are fast becoming the place to share and discover news, activities, and content of interest. Facebook alone reports more than 1 billion users, each having on average 130 friends and connected to 80 communities, and spending on Facebook less than one hour a day. The volume of the generated content of potential interest is, thus, overwhelming and ever growing, but the time spent on the social networks is fairly limited.\u0000 How can users stay abreast of the activities of interest given this severe information overload? Activity feed is a simple mechanism deployed nowadays by many social networks, which performs information filtering on the users' behalf. Typically, activity feed encompasses reverse chronologically ordered items corresponding to activities carried out by direct friends and followees. However, activity feed can hardly cope with the volume and diversity of the activities. In order to alleviate information overload, simplify content discovery, and sustain user engagement, there is a need to personalise the activity feed, i.e., identify items of a particular interest and relevance for the user and filter out irrelevant items.\u0000 The feed personalisation task can be naturally represented as a top-K recommendation problem. Let us denote by N the set of items corresponding to activities that can potentially be included in the feed, e.g., all the activities carried out since the user's last visit. Hence, the personalisation task aims at selecting and recommending a smaller set of items, K (|K|≪|N|), corresponding to activities of the highest relevance for the user. Essentially, the recommendation process entails scoring all the |N| candidate items and selecting |K| top-scoring items.\u0000 What information can facilitate the item scoring? When interacting with a social network, users typically leave very little explicit feedback, primarily their 'likes'. There is a moderate amount of strong implicit user-to-user feedback, e.g., friending and direct communication (messages and comments), and abundance of weak implicit user-to-activity feedback, such as content viewing and contribution, community membership, and event participation. Finally, there is some self-reported and often unreliable information pertaining to user demographics, location, preferences, skills, or interests.\u0000 How can all this this information be modelled, fused, mined, and eventually leveraged for scoring and recommending activity feed items? This problem has been investigated from different angles in the recent years [1-10]. In this talk, we will overview most prominent works into the personalisation of the activity feed. These works proposed a spectrum of algorithmic approaches and evaluated them with numerous social networks of a highly heterogeneous nature. We will summarise the main components","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122876701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Twitter is a popular medium for discussing unfolding events in real-time. Due to the large volume of user generated data during these events, it's important to be able recommend the best content while it's fresh. Current recommendation algorithms for Twitter take into account the user's tweets and her social network, but since real-time events might be unique or unexpected, the history of a user may not be sufficient for finding the most relevant content. Additionally, for users who want to join the conversation at that specific moment (or follow it without having to create an account), the system will be faced with the cold-start problem. We propose a simple visualization technique that considers the activity of the whole community participating in the real-time discussion, by capturing their co-retweeting behavior. Such a technique depicts the big picture, allowing a user to choose content from parts of the community that share her opinions or beliefs.
{"title":"Visualizing co-retweeting behavior for recommending relevant real-time content","authors":"Samantha Finn, Eni Mustafaraj","doi":"10.1145/2463656.2463660","DOIUrl":"https://doi.org/10.1145/2463656.2463660","url":null,"abstract":"Twitter is a popular medium for discussing unfolding events in real-time. Due to the large volume of user generated data during these events, it's important to be able recommend the best content while it's fresh. Current recommendation algorithms for Twitter take into account the user's tweets and her social network, but since real-time events might be unique or unexpected, the history of a user may not be sufficient for finding the most relevant content. Additionally, for users who want to join the conversation at that specific moment (or follow it without having to create an account), the system will be faced with the cold-start problem. We propose a simple visualization technique that considers the activity of the whole community participating in the real-time discussion, by capturing their co-retweeting behavior. Such a technique depicts the big picture, allowing a user to choose content from parts of the community that share her opinions or beliefs.","PeriodicalId":136302,"journal":{"name":"MSM '13","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114937700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：