Abstract
One of the challenges in fashion recommendations is how to incorporate the concepts of fashion and style to provide a more tailored personalized experience for fashion lovers. Despite that these concepts are subjective, our fashion experts at Farfetch have defined a few key sets of aesthetics which attempt to capture the essence of users’ styles into groups. This categorization will help us to understand the customers’ fashion preferences and hence guide our recommendations through the subjectivity. In this paper, we will demonstrate that such concepts can be predicted from users’ behaviors and the products they have interacted with. We not only compared a popular machine learning algorithm—Random Forest with a more recent deep learning algorithm—Convolutional Neural Network (CNN), but also looked at 3 different sets of features: text, image, and inferred user statistics, together with their various combinations in building such models. Our results show that it is possible to identify a customer’s aesthetic based on this data. Moreover, we found that the use of the textual descriptions of products interacted by the customer led to better classification results.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
It builds a single generalized model capable of processing output correlations. To build a tree, it uses a multi-output splitting criteria computing average impurity reduction across all the outputs. To the best of our knowledge, this could be viewed as a kind of greedy label powerset technique.
References
Bojanowski P, Grave E, Joulin A, Mikolov T (2016) Enriching word vectors with subword information. arXiv preprint arXiv:1607.04606 (2016)
Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP (2020) Smote: Synthetic minority over-sampling technique. arXiv preprint arXiv:1106.1813 (2020)
Dembczynski K, Cheng W, Hüllermeier E (2010) Bayes optimal multilabel classification via probabilistic classifier chains. In: Fürnkranz J, Joachims T (eds) ICML, pp 279–286. Omnipress (2010). http://dblp.uni-trier.de/db/conf/icml/icml2010.html#DembczynskiCH10
Faggella D (2017) The roi of recommendation engines for marketing. https://martechtoday.com/roi-recommendation-engines-marketing-205787 (2017). Accessed 29 July 2020
Gonçalves D, Liu L, Magalhães A (2019) How big can style be? addressing high dimensionality for recommending with style. CoRR abs/1908.10642. http://dblp.uni-trier.de/db/journals/corr/corr1908.html#abs-1908-10642
He K, Zhang X, Ren S, Sun J (2015) Deep residual learning for image recognition. arXiv preprint arXiv:1512.03385
Hsiao WL, Grauman K (2017) Learning the latent “look”: Unsupervised discovery of a style-coherent embedding from fashion images. In: ICCV, pp 4213–4222. IEEE Computer Society. http://dblp.uni-trier.de/db/conf/iccv/iccv2017.html#HsiaoG17
Iqbal M, Aryafar K, Anderton T (2019) Style conditioned recommendations. CoRR abs/1907.12388. http://dblp.uni-trier.de/db/journals/corr/corr1907.html#abs-1907-12388
Lee H, Seol J, Lee S (2019) Style2vec: Representation learning for fashion items from style sets. CoRR abs/1708.04014. http://arxiv.org/abs/1708.04014
Liu F, Zhang X, Ye Y, Zhao Y, Li Y (2015) Mlrf: Multi-label classification through random forest with label-set partition. In: Huang DS, Han K (eds) Advanced intelligent computing theories and applications. Springer International Publishing, Cham, pp 407–418
Liu J, Chang WC, Wu Y, Yang Y (2017) Deep learning for extreme multi-label text classification. pp 115–124. https://doi.org/10.1145/3077136.3080834
M. Hadi Kiapour Kota Yamaguchi ACBTLB (2014) Hipster wars: Discovering elements of fashion styles. In: European conference on computer vision (2014)
Magalhães AR (2019) The trinity of luxury fashion recommendations: data, experts and experimentation. In: Bogers T, Said A, Brusilovsky P, Tikk D (eds) RecSys, p 522. ACM. http://dblp.uni-trier.de/db/conf/recsys/recsys2019.html#Magalhaes19
Mahajan D, Girshick R, Ramanathan V, He K, Paluri M, Li Y, Bharambe A, van der Maaten L (2018) Exploring the limits of weakly supervised pretraining. arXiv preprint arXiv:1805.00932
Mahmood A, Ospina AG, Bennamoun M, An S, Sohel F, Boussaid F, Hovey R, Fisher RB, Kendrick G (2020) Automatic hierarchical classification of kelps using deep residual features. arXiv preprint arXiv:1906.10881
Mikolov T, Chen K, Corrado G, Dean J (2013) Efficient estimation of word representations in vector space, 1–12. http://arxiv.org/abs/1301.3781
Nam J, Kim J, Loza Mencía E, Gurevych I, Fürnkranz J (2014) Large-scale multi-label text classification – revisiting neural networks. In: Calders T, Esposito F, Hüllermeier E, Meo R (eds) Machine learning and knowledge discovery in databases. Springer, Berlin Heidelberg, Berlin, Heidelberg, pp 437–452
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay E (2011) Scikit-learn: Machine learning in Python. Journal of machine learning research 12:2825–2830
Pennington J, Socher R, Manning CD (2014) Glove: global vectors for word representation. In: Empirical methods in natural language processing (EMNLP), pp 1532–1543. http://www.aclweb.org/anthology/D14-1162
Simo-Serra E, Ishikawa H (2016) Fashion style in 128 floats: Joint ranking and classification using weak data for feature extraction. In: CVPR, pp 298–307. IEEE computer society (2016). http://dblp.uni-trier.de/db/conf/cvpr/cvpr2016.html#Simo-SerraI16
Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556
Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z (2015) Rethinking the inception architecture for computer vision. arXiv preprint arXiv:1512.00567
Veit A, Kovacs B, Bell S, McAuley JJ, Bala K, Belongie SJ (2015) Learning visual clothing style with heterogeneous dyadic co-occurrences. In: ICCV, pp 4642–4650. IEEE computer society. http://dblp.uni-trier.de/db/conf/iccv/iccv2015.html#VeitKBMBB15
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Liu, L., Silva, I., Nogueira, P., Magalhães, A., Martins, E. (2021). User Aesthetics Identification for Fashion Recommendations. In: Dokoohaki, N., Jaradat, S., Corona Pampín, H.J., Shirvany, R. (eds) Recommender Systems in Fashion and Retail. Lecture Notes in Electrical Engineering, vol 734. Springer, Cham. https://doi.org/10.1007/978-3-030-66103-8_3
Download citation
DOI: https://doi.org/10.1007/978-3-030-66103-8_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-66102-1
Online ISBN: 978-3-030-66103-8
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)