--======== Review Reports ========-- The review report from reviewer #1: *1: Is the paper relevant to ICDM? [_] No [X] Yes *2: How innovative is the paper? [_] 6 (Very innovative) [_] 3 (Innovative) [X] -2 (Marginally) [_] -4 (Not very much) [_] -6 (Not at all) *3: How would you rate the technical quality of the paper? [_] 6 (Very high) [X] 3 (High) [_] -2 (Marginal) [_] -4 (Low) [_] -6 (Very low) *4: How is the presentation? [_] 6 (Excellent) [X] 3 (Good) [_] -2 (Marginal) [_] -4 (Below average) [_] -6 (Poor) *5: Is the paper of interest to ICDM users and practitioners? [X] 3 (Yes) [_] 2 (May be) [_] 1 (No) [_] 0 (Not applicable) *6: What is your confidence in your review of this paper? [_] 2 (High) [X] 1 (Medium) [_] 0 (Low) *7: Overall recommendation [_] 6: must accept (in top 25% of ICDM accepted papers) [X] 3: should accept (in top 80% of ICDM accepted papers) [_] -2: marginal (in bottom 20% of ICDM accepted papers) [_] -4: should reject (below acceptance bar) [_] -6: must reject (unacceptable: too weak, incomplete, or wrong) *8: Summary of the paper's main contribution and impact In this paper, a method with name Fossil fusing similarity based recommendation method with Markov chains is proposed to make personalized sequential recommendation. With this method, both long-term preference and short-term dynamics are taken into account. Meanwhile sparsity issue is also considered. Experiments were conducted to evaluate the performance of Fossil and the results demonstrated improvement of recommendation accuracy. *9: Justification of your recommendation I recommend this paper for acceptance for three reasons. 1. The idea considering both long-term user preferences and short-term sequential dynamics is reasonable and important for effective recommendation. 2. The proposed method is novel with profound foundations in existing literatures. 3. The technical presentation is fluent. *10: Three strong points of this paper (please number each point) 1) A new method fusing similarity based recommendation method with Markov chains is proposed 2) Experiments were conducted to evaluate the performance of the proposed method. 3) Writing is clear. *11: Three weak points of this paper (please number each point) 1) Lack analysis of time complexity of the proposed method. 2) Discussion of the applicability of the proposed method is missing. 3) The improvement compared with existing methods is small. *12: Is this submission among the best 10% of submissions that you reviewed for ICDM'16? [X] No [_] Yes *13: Would you be able to replicate the results based on the information given in the paper? [X] No [_] Yes *14: Are the data and implementations publicly available for possible replication? [X] No [_] Yes *15: If the paper is accepted, which format would you suggest? [X] Regular Paper [_] Short Paper *16: Detailed comments for the authors In this paper, a method with name Fossil fusing similarity based recommendation method with Markov chains is proposed to make personalized sequential recommendation. With this method, both long-term preference and short-term dynamics are taken into account. Meanwhile sparsity issue is also considered. Experiments were conducted to evaluate the performance of Fossil and the results demonstrated improvement of recommendation accuracy. Generally speaking, the methodology is novel, though the application is not novel. Also, the proposed method has profound foundations in literatures, especially by borrowing ideas from recommendation models FISM and FPMC. At the same time, I have following concerns. 1. A theoretical complexity analysis is needed. Also, an experimental comparison of execution efficiency of different methods is necessary. 2. More evaluation metrics such as Recall, Precision, or NDCG are needed. Since all methods directly optimize AUC on the training set and this research only presents the results in terms of AUC, the slight improvement of Fossil may be caused by over-fitting. 3. The performance improvement is small. ======================================================== The review report from reviewer #2: *1: Is the paper relevant to ICDM? [_] No [X] Yes *2: How innovative is the paper? [_] 6 (Very innovative) [X] 3 (Innovative) [_] -2 (Marginally) [_] -4 (Not very much) [_] -6 (Not at all) *3: How would you rate the technical quality of the paper? [_] 6 (Very high) [X] 3 (High) [_] -2 (Marginal) [_] -4 (Low) [_] -6 (Very low) *4: How is the presentation? [_] 6 (Excellent) [X] 3 (Good) [_] -2 (Marginal) [_] -4 (Below average) [_] -6 (Poor) *5: Is the paper of interest to ICDM users and practitioners? [X] 3 (Yes) [_] 2 (May be) [_] 1 (No) [_] 0 (Not applicable) *6: What is your confidence in your review of this paper? [X] 2 (High) [_] 1 (Medium) [_] 0 (Low) *7: Overall recommendation [X] 6: must accept (in top 25% of ICDM accepted papers) [_] 3: should accept (in top 80% of ICDM accepted papers) [_] -2: marginal (in bottom 20% of ICDM accepted papers) [_] -4: should reject (below acceptance bar) [_] -6: must reject (unacceptable: too weak, incomplete, or wrong) *8: Summary of the paper's main contribution and impact The paper focuses on user preference as dynamics, which consists of long-term and short-term ones. Then, the authors define the objective function of sum of those effects which should be maximized. *9: Justification of your recommendation From the sequential point of views, the authors assume two aspects of preferences: converged one and fluctuated one. Both have different effects on decision making. *10: Three strong points of this paper (please number each point) Sequential recommendation with two major temporal dynamics Definition of objective functions and maximization algorithm Experimental evaluation *11: Three weak points of this paper (please number each point) Assumption: Is long-term dynamics really a converged one ? Data sensitivity: the above assumption may impose a restriction on data analysis framework. Convergence speed: Is the proposed method faster than others ? *12: Is this submission among the best 10% of submissions that you reviewed for ICDM'16? [_] No [X] Yes *13: Would you be able to replicate the results based on the information given in the paper? [X] No [_] Yes *14: Are the data and implementations publicly available for possible replication? [_] No [X] Yes *15: If the paper is accepted, which format would you suggest? [X] Regular Paper [_] Short Paper *16: Detailed comments for the authors The paper gives an interesting view on user preference. The long-term, and short-term dynamics may be referred to global and local dynamics in the ordinary dynamics or geometrical point of view. I have one question. The overall method assumes that long-term dynamics will be converged. Otherwise, short-term dynamics gives a different meaning to the context of user preference. It may be a subtle problem and difficult to answer, but this kind of assumption may give some restriction to the proposed method. ======================================================== The review report from reviewer #3: *1: Is the paper relevant to ICDM? [_] No [X] Yes *2: How innovative is the paper? [_] 6 (Very innovative) [_] 3 (Innovative) [_] -2 (Marginally) [X] -4 (Not very much) [_] -6 (Not at all) *3: How would you rate the technical quality of the paper? [_] 6 (Very high) [_] 3 (High) [X] -2 (Marginal) [_] -4 (Low) [_] -6 (Very low) *4: How is the presentation? [_] 6 (Excellent) [_] 3 (Good) [X] -2 (Marginal) [_] -4 (Below average) [_] -6 (Poor) *5: Is the paper of interest to ICDM users and practitioners? [X] 3 (Yes) [_] 2 (May be) [_] 1 (No) [_] 0 (Not applicable) *6: What is your confidence in your review of this paper? [X] 2 (High) [_] 1 (Medium) [_] 0 (Low) *7: Overall recommendation [_] 6: must accept (in top 25% of ICDM accepted papers) [X] 3: should accept (in top 80% of ICDM accepted papers) [_] -2: marginal (in bottom 20% of ICDM accepted papers) [_] -4: should reject (below acceptance bar) [_] -6: must reject (unacceptable: too weak, incomplete, or wrong) *8: Summary of the paper's main contribution and impact This paper proposes an innovative approach to address the problem of personalized sequential recommendation. The authors take into account both long-term user preference and short-term dynamics, and develop a unified method by combining similarity-based method (FISM) with Markov Chains. To validate the effectiveness of the proposed method, they evaluated their method on a variety of large, real-world datasets, and provided qualitative analysis as well. *9: Justification of your recommendation This work aim to tackle the problem of sequential personalized recommendation, which is an interesting and will provide valuable insights for practitioners. The evaluation in this paper is solid, the idea is simple and the presentation is clear. *10: Three strong points of this paper (please number each point) 1.This work provides extensive empirical study to evaluate the performance of the proposed method. 2.The presentation is clear and not hard to follow. 3.They provide the additional visualization results to demonstrate the importance of both long-term preference and short-term dynamics for personalized recommendations. *11: Three weak points of this paper (please number each point) 1.The idea is not innovative. 2.The technical contribution is limited. 3.There exist typos and grammatical errors in the presentation, as well as typos in equation (5), (8) and (9). *12: Is this submission among the best 10% of submissions that you reviewed for ICDM'16? [X] No [_] Yes *13: Would you be able to replicate the results based on the information given in the paper? [X] No [_] Yes *14: Are the data and implementations publicly available for possible replication? [X] No [_] Yes *15: If the paper is accepted, which format would you suggest? [_] Regular Paper [X] Short Paper *16: Detailed comments for the authors This work aims to develop a combinative method (Fossil) for personalized sequential recommendation by integrating Factored Item Similarity Models (FISM) and Markov Chains. To demonstrate the effectiveness of their proposed method, the authors conduct extensive experiments on a variety of large, real-world datasets. The experiments show that Fossil outperforms the baselines. However, there are several disadvantages in this paper that should be overcame before publication First, they clarify that their method “significantly outperforms alternative algorithms” in the Abstract, but the improvements comparing with baseline methods are not too much (according to Table IV). This should be modified. Second, the technical contribution of this paper is limited. The basic idea of their proposed method, Fossil, is similar to FPMC. To alleviate the data sparsity issue, this paper exploits the FISM method and replaces MF module in FPMC with FISM. Whereas, the quantitative comparison evaluation regarding the data sparsity is not provided. Third, improve the writing and overall presentation of the paper. The authors must in particular make sure their paper is free from grammatical errors. ========================================================