Previous sequential-recommendation methods have been able to capture patterns of item characteristics that interact with the user. However, they modeled user behavior using a whole interaction sequence, despite possible changes in a user's behavior over time, which can make some behaviors no longer relevant by the end of the user-item interaction period. That is, each item representation was derived from the influences of all items in the whole sequence (i.e. global representation) without considering item-adjacency factors that can affect the item characteristics during the interaction period (i.e. local representation). Furthermore, these methods modeled only user behavior, ignoring item behavior, which can involve patterns of user characteristics for users who interact with the item. In this paper, we therefore propose a novel attentive local-interaction model for sequential recommendation called ARERec, which applies a region-embedding technique to both user and item historical sequences. The aim is to model user and item behavior over the user-item interaction sequence while considering local representations that contain specific characteristics of both user and item in the sequence. In this way, information is derived for corresponding periods that reflect more-specific reasons behind the interaction in the sequence. Moreover, to account for ratings and neighborrelated factors, we adopt the concept of neighbor-based collaborative filtering in our predictions. One issue is that neighbors have the same similarity levels for all users, resulting in similar predictions, even though there may be different user-specific information. To address this problem, we apply a multi-head attention mechanism to personalize each neighbor based on the user's characteristics. Extensive experiments on three datasets demonstrate that ARERec consistently outperforms state-of-the-art sequential methods, including Recurrent Neural Networks and the attention-based methods (both unidirectional and bidirectional) HitRate and Normalized Discounted Cumulative Gain. Our experiments also show that ARERec provides superior results by considering interaction periods with local representation rather than the whole sequence using global representation.

Previous personalized hashtag recommendations have been able to recommend suitable hashtags for a given microblog. Despite their performance improvement, we argue that three challenges remain unexplored. First, prior studies capture user interests solely from user-hashtag interactions that are directly connected (i.e., first-order relations), making them unable to deal with multiple user behaviors, including user-user social and hashtag-hashtag co-occurrence, and also restrict relations from similar users that are indirectly connected (i.e., high-order relations). Second, previous works personalize content at the microblog level, ignoring the personalized aspects that users have for each word in the microblog. Third, past studies capture correlations among hashtags in the same microblog from only the left-side correlations, restricting the right-side correlations. In this paper, we propose a novel integral model for personalized hashtag recommendation named PAC-MAN, which explores high-order multiple relations to model fruitful user and hashtag representation before fusing with word representation for word-level personalization and integrating with sequenceless hashtag correlation for the recommendation. First, to derive fruitful user and hashtag representation from higher-order multiple relations, Multi-relational Attentive Network (MAN) applies GNN to jointly capture relations on three communities: (1) user-hashtag interaction; (2) user-user social; and (3) hashtag-hashtag co-occurrence. Second, to personalize content at a word level, Person-And-Content based BERT (PAC) extends BERT to input not only word representations from the microblog but also the fruitful user representation from MAN, allowing each word to be fused with user aspects. Finally, to capture sequenceless hashtag correlations, the fruitful hashtag representations from MAN that contain the hashtag's community perspectives are inserted into BERT to integrate with the hashtag's wordsemantic perspectives, and a hashtag prediction task is then conducted under the mask concept, enabling hashtag correlations to be obtained from both left and right sides without sequence constraints. Extensive experiments on the Twitter dataset demonstrate that PAC-MAN consistently outperforms state-of-the-art methods, including neural network based and traditional graph based methods, over precision, recall, and F1-score metrics.