1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) (Regular Paper) 24 3,

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) (Regular Paper) 24 3, 2019 5 (JBE Vol. 24, No. 3, May 2019) https://doi.org/10.5909/jbe.2019.24.3.515 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a), a) Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement Hyeon-woo An a) and Nammee Moon a)....,. Abstract In the real world, the user's preference for a particular product is determined by many factors besides the quality of the product. The reflection of these external factors was very difficult because of various fundamental problems including lack of data. However, access to external factors has become easier as the infrastructure for public data is opened and the availability of evaluation platforms with diverse and vast amounts of data. In accordance with these changes, this paper proposes a recommendation system structure that can reflect the collectable factors that affect user's preference, and we try to observe the influence of actual influencing factors on preference by applying case. The structure of the proposed system can be divided into a process of selecting and extracting influencing factors, a process of supplementing insufficient data using sentence analysis, and finally a process of combining and merging user's evaluation data and influencing factors. We also propose a validation process that can determine the appropriateness of the setting of the structural variables such as the selection of the influence factors through comparison between the result group of the proposed system and the actual user preference group. Keyword : Hybrid Recommendation, influencing Factor, Recommendation System Copyright 2016 Korean Institute of Broadcast and Media Engineers. All rights reserved. This is an Open-Access article distributed under the terms of the Creative Commons BY-NC-ND (http://creativecommons.org/licenses/by-nc-nd/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited and not altered.

(JBE Vol. 24, No. 3, May 2019). SNS.,,....,..,....,. a) (Department of Computer Engineering, Hoseo University) Corresponding Author : (Nammee Moon) E-mail: nammee.moon@gmail.com Tel: +82-2-2059-2310 ORCID: http://orcid.org/0000-0003-2229-4217 This work has supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2017R1A2B40 08886). 2019 ( ) (No. NRF-2017R1A2B4008886). Manuscript received April 30, 2019; Revised May 16, 2019; Accepted May 16, 2019..,,,.., 4 (,,, ) 5 (,,,, ) 20.,. [1]..., (IFBHR : Influence Factor Based Hybrid Recommendation system), IFBHR,.. 1..

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) [2]. (first rater problem) [3].,,., [4].. (first rater) (Over Specialization).. 2. (first rater).. [5]..,... [6]... 3. IFBHR. LDA (Latent dirichlet Allocation),, [7]..,,, [8].. IFBHR..

(JBE Vol. 24, No. 3, May 2019) 1. ( :, :,, X:, ) Table 1. Similar system comparison chart( : possible, existence : partial possible, partial existence, X: impossible, not present)) LDA-based article recommendation [7] Multi-profile-based music recommendation [8] Deep auto-encoder based personalized recommendation [9] IFBHR Source of influence factor External Individual / External Individual External Recommendation type Hybrid Hybrid CF(Collaborative Filtering) Hybrid Utilization Learning Model STPM (Spatial Topical Preference Model) Multiple regression analysis SDAE (Stacked Denoising AutoEncoder) Appropriate model selection can be applied according to supplementary data Recommended results article music genre rating score preference Personalized recommendations Evaluation data supplement X X X No user evaluation history required X X Solving Over Specialization Problems application expandable X X Relative operation speed (Required operation) normal (location estimation, topic calculation) fast (Profile acquisition) Slow (User review extraction) fast (Extraction of influence factor)., Stacked Denoising Auto- Encoder (SDAE) [9]. 1 IFBHR.. IFBHR(Influence Factor Based Hybrid Recommendation-system) IFBHR(Influence Factor Based Hybrid Recommendationsystem).,..,.., ( ).. 1. 1. phase 1 :. (API, ).......

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) 1. IFBHR Fig. 1. The overall flow diagram of the IFBHR,. phase 2 :..... phase 3 :. 2.,.. 2..,. Yi :, k X1i, X2i.. Xki : Yi = α + β1 X1i + β2 X2i +. + βk Xki + ε. Y

(JBE Vol. 24, No. 3, May 2019) (i>=k).,,.,,. [10,11].. [12],..,... S, I, S.c E.c I. S E I : I = (S) (S.c=Ei.c)(E),... API.,.,. ( API, ) (,, ). API. 2. 2. Fig. 2. Factor extraction process 3..,.,..

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement).,.,.. IFBHR CNN... 3..... S:, P:, m: : 4. 3. Fig. 3. Data Supplementation Process 4. /...., 4. Fig. 4. Weighted average sum pseudo code. /

(JBE Vol. 24, No. 3, May 2019). 5..,. MAE(Mean Absolute Error). 20. 5. 5. Fig. 5. Verification process 1.... 197. 2.,..,,..,. python selenium TripAdvisor. 49089....,,., 170 TripAdvisor 111,652. 3. TripAdvisor

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) Cutting Konlpy Okt( Twitter),.. CNN(Convolutional Neural Network) [13] 1~5.,. 4 (,,, ),,... lat,lon (E:i, S: ) : lati, loni = π(s.lat,s.lon)(s) (S.name=Ei.name)(Ei) 7. Fig. 7. Obtain the nearest station Pseudo code (E.date) (W: ) : weather=(ei) W.date=Ei.date(σ(W.obs_id=Closest_ obsi)(w)) 8. Fig. 8. Influence factors acquisition pseudo code 2. 2. Table 2. Merge criteria table 6. Fig. 6. Pseudo code to acquire tourist spot coordinates (O.id) (O:, dist(lat1, lon1, lat2, lon2): ) : Closest_obsi = π(o.id) (DNO O.id MIN(id) KEEP ( DENSE_RANK LAST ORDER BY dist(o.lat,o.lon,lati,loni) DESC )O)

(JBE Vol. 24, No. 3, May 2019). 9. [,, 1, 2,,,, [lat, lon], ]. 9. Fig. 9. Rating table and merge process output 5 0.406 10 0.3011, 20 0.2154.,. 64 10 95,207 10, 11. 4. 2 CNN 3. 3. Table 3. Experimental use learning parameter parameter value Contents embedding_dim 32 Dimension of Embedding Word Vector filter_sizes (3,4,5) Size of filter. It acts like the kernel in image analysis. num_filters 128 Number of convolution channels dropout_keep_prob 0.2 l2_reg_lambda 0.2 It deals with the weight of neurons to be learned during learning. It is possible to prevent over-fitting. The lambda value of the l2 normalization. The degree of normalization can be adjusted. 10. 10 Fig. 10. Data Distribution Ratio for the Top 10 Categories 64 6000 53%, 54%. 822,547. 3 4 1. MAE (Mean Absolute Error). 11. 10 Fig. 11. Affinity charts for the top 10 categories S1 44% ( S3',22%), ( S0',17%), ( S2',17%).

1 : (Hyeon-woo An et al.: Influential Factor Based Hybrid Recommendation System with Deep Neural Network-Based Data Supplement) S1 'B1'. 8 S3C0B1'( ).... IFBHR,,. (References) [1] J. Son, S. Kim, H. Kim and S. Cho. "Review and Analysis of Recommender Systems" Journal of the Korean Institute of Industrial Engineers, Vol. 41, No. 2, pp. 185-208, April 2015, https://doi.org/10.7232/jkiie.2015.41.2.185 (accessed April. 15, 2015). [2] Goldberg, David, et al. "Using collaborative filtering to weave an information Tapestry." Communications of the ACM, Vol. 35, No. 12, pp. 61-71, Dec 1992. (https://go.galegroup.com/ps/anonymous?id=gale%7ca13039895 &sid=googlescholar&v=2.1&it=r&linkaccess=abs&issn =00010782&p=AONE&sw=w) [3] Su, Xiaoyuan, and Taghi M. Khoshgoftaar. "A survey of collaborative filtering techniques." Advances in artificial intelligence, Vol. 2009, Article ID 421425, 19 pages, 2009, https://doi.org/10.1155/2009/ 421425 (accessed Aug. 3, 2009). [4] Wu, Yi-Hung, and Arbee LP Chen. "Index structures of user profiles for efficient web page filtering services." Proceedings 20th IEEE International Conference on Distributed Computing Systems. IEEE, April 2000. (DOI. 10.1109/ICDCS.2000.840981) [5] Adomavicius, Gediminas, and Alexander Tuzhilin. "Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions." IEEE Transactions on Knowledge & Data Engineering 6. vol. 17, pp. 734-749, June 2005. (DOI. 10.1109/TKDE.2005.99) [6] Soboroff, Ian, and Charles Nicholas. "Combining content and collaboration in text filtering." Proceedings of the IJCAI. Vol. 99. pp. 86-91, sn, 1999. (https://www.csee.umbc.edu/csee/research/cadip/1999symposium/ mlif.pdf) [7] Noh, Yunseok, Yong-Hwan Oh, and Seong-Bae Park. "A locationbased personalized news recommendation." 2014 International Conference on Big Data and Smart Computing (BIGCOMP). IEEE, 2014. (DOI. 10.1109/BIGCOMP.2014.6741416) [8] Park, Kyong-Su, and Nam-Me Moon. "Multidimensional Optimization Model of Music Recommender Systems." The KIPS Transactions: PartB. Vol. 19, No. 3, pp. 155-164, June 2012, https://doi.org/10.3745/ KIPSTB.2012.19B.3.155 (accessed Feb. 31, 2012) [9] Hyunwoo Je, Junwoo Kim, Mun Y. Yi. "Deep AutoEncoder based Personalized Recommendation System : Considering user s intrinsic characteristics." KOREA INFORMATION SCIENCE SOCIETY. 773-775. June 2017. (http://www.dbpia.co.kr/journal/articledetail?nodeid=node 07207377&language=ko_KR) [10] Scott, D., and Chr Lemieux. "Weather and climate information for tourism." Procedia Environmental Sciences. Vol 1, pp. 146-183, 2010, https://doi.org/10.1016/j.proenv.2010.09.011 (accessed Nov. 18, 2010) [11] Becken, Susanne, and Jude Wilson. "The impacts of weather on tourist travel." Tourism Geographies. Vol. 15, No. 4, pp. 620-639, Feb 2013, https://doi.org/10.1080/14616688.2012.762541 (accessed Feb. 12, 2013) [12] Dzogang, Fabon, Stafford Lightman, and Nello Cristianini. "Diurnal variations of psychometric indicators in Twitter content." PloS one. Vol. 13, No. 6, e0197002, June 2018 (https://journals.plos.org/plosone/article/file?id=10.1371/ journal.pone.0197002&type=printable) [13] Kim, Yoon. "Convolutional neural networks for sentence classification." arxiv preprint arxiv:1408.5882, Aug 2014. (https://arxiv.org/abs/1408.5882)

(JBE Vol. 24, No. 3, May 2019) - 2018 : - 2018 ~ : - ORCID : https://orcid.org/0000-0003-2880-5639 - :,, (AI) - 1985 : - 1987 : - 1998 : - 1999 ~ 2003 : - 2003 ~ 2008 : - 2008 ~ : - ORCID : http://orcid.org/0000-0003-2229-4217 - : Social Learning,, HCI,, User Centric data analysis