• Corpus ID: 20443433

An LSTM-Based Dynamic Customer Model for Fashion Recommendation

  title={An LSTM-Based Dynamic Customer Model for Fashion Recommendation},
  author={Sebastian Heinz and Christian Bracher and Roland Vollgraf},
Online fashion sales present a challenging use case for personalized recommendation: Stores offer a huge variety of items in multiple sizes. Small stocks, high return rates, seasonality, and changing trends cause continuous turnover of articles for sale on all time scales. Customers tend to shop rarely, but often buy multiple items at once. We report on backtest experiments with sales data of 100k frequent shoppers at Zalando, Europe's leading online fashion platform. To model changing customer… 

Figures and Tables from this paper

Aspect-Based Fashion Recommendation With Attention Mechanism
An aspect-based fashion recommendation model with attention mechanism (AFRAM) to predict customer ratings based on online reviews of fashion products is proposed and demonstrated that AFRAM is more effective in customer rating predictions, as compared to several state-of-the-art fashion recommenders.
Disentangling Features for Fashion Recommendation
This paper addresses the task of compatible garment recommendation developing a MANN architecture by taking into account the co-occurrence of clothing attributes, such as shape and color, to compose an outfit by obtaining disentangled representations of fashion items and storing them in external memory modules.
Towards Preprocessing Guidelines for Neural Network Embedding of Customer Behavior in Digital Retail
This paper provides a set of guidelines for preprocessing and modeling consumer transaction data to learn neural network embeddings, and introduced a multi-task Long Short-Term Memory Network to evaluate the guidelines proposed through the task of purchase behavior prediction.
Fashion Recommendation Systems, Models and Methods: A Review
This paper is the first scholarly article to review the state-of-the-art fashion recommendation systems and the corresponding filtering techniques and explores various potential models that could be implemented to developfashion recommendation systems in the future.
A Review of Modern Fashion Recommender Systems
The most pressing challenges in fashion RS research are identified and a taxonomy that categorizes the literature according to the objective they are trying to accomplish is created and the most commonly used datasets and evaluation metrics are created.
Reinforced KGs reasoning for explainable sequential recommendation
This work proposes a novel architecture called Reinforced Sequential Learning with Gated Recurrent Unit (RSL-GRU), which is composed of a Reinforced Path Reasoning Network (RPRN) component and a GRU component that takes users’ sequential behaviors and their associated KGs in chronological order as input and outputs potential top-N items from a global perspective.
Outfit Generation and Style Extraction via Bidirectional LSTM and Autoencoder
An unsupervised style extraction module is incorporated into a model to learn outfits using the style information of an outfit as a whole to generate outfits more flexibly without requiring additional information.
This article aims to provide a comprehensive review of recent research efforts on deep learning based clothing matching and categorized deep learning-based clothing matching methods into two categories, namely collaborative method and content-based method.
Is Language Modeling Enough? Evaluating Effective Embedding Combinations
It is observed that adding topic model based embeddings helps for most tasks and that differing pre-training tasks encode complementary features, and new state of the art results on the MPQA and SUBJ tasks in SentEval are presented.
Smart Fashion: A Review of AI Applications in the Fashion & Apparel Industry
A structured task-based multi-label classification of fashion research articles provides researchers with explicit research directions and facilitates their access to the related studies, improving the visibility of studies simultaneously.


Collaborative Recurrent Neural Networks for Dynamic Recommender Systems
This work proposes a novel, flexible and expressive collaborative sequence model based on recurrent neural networks that captures a user’s contextual state as a personalized hidden vector by summarizing cues from a data-driven, thus variable, number of past time steps, and represents items by a real-valued embedding.
Fashion DNA: Merging Content and Sales Data for Recommendation and Article Mapping
We present a method to determine Fashion DNA, coordinate vectors locating fashion items in an abstract space. Our approach is based on a deep neural network architecture that ingests curated article
Long and Short-Term Recommendations with Recurrent Neural Networks
This work characterized the full short-term/long-term profile of many collaborative filtering methods, and showed how recurrent neural networks can be steered towards better short or long-term predictions, showing that RNNs are perfectly suited for collaborative filtering on dense datasets where it outperforms traditional item recommendation algorithms.
Collaborative Recurrent Autoencoder: Recommend while Learning to Fill in the Blanks
Experiments on real-world datasets from different domains show that, by jointly modeling the order-aware generation of sequences for the content information and performing CF for the ratings, CRAE is able to significantly outperform the state of the art on both the recommendation task based on ratings and the sequence generationtask based on content information.
Logistic Matrix Factorization for Implicit Feedback Data
Logistic MF is presented, a new probabilistic model for matrix factorization with implicit feedback that is simple to implement, highly parallelizable, and has the added benefit that it can model the probability that a user will prefer a specific item.
Temporal Collaborative Filtering with Bayesian Probabilistic Tensor Factorization
This work proposes a factor-based algorithm that is able to take time into account, and provides a fully Bayesian treatment to avoid tuning parameters and achieve automatic model complexity control.
Long Short-Term Memory
A novel, efficient, gradient based method called long short-term memory (LSTM) is introduced, which can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units.
Matrix Factorization Techniques for Recommender Systems
As the Netflix Prize competition has demonstrated, matrix factorization models are superior to classic nearest neighbor techniques for producing product recommendations, allowing the incorporation of
Optimizing Classifier Performance via an Approximation to the Wilcoxon-Mann-Whitney Statistic
This work proposes an objective function that is an approximation to the Wilcoxon-Mann-Whitney statistic, which is equivalent to the AUC, and applies it to real-world customer behavior prediction problems for a wireless service provider and a cable service provider, and achieves reliable improvements in the ROC curve.
Deep Residual Learning for Image Recognition
This work presents a residual learning framework to ease the training of networks that are substantially deeper than those used previously, and provides comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.