Figure - available from: SN Computer Science
This content is subject to copyright. Terms and conditions apply.
Source publication
Taxonomies are used as product categories to facilitate users navigating through an e-commerce portal with the help of hierarchically structured concepts. However, the identical taxonomy is shown to each customer regardless of the channel used. This is challenging for the customers in terms of user experience, as the screen size is rigid, and has n...
Citations
... For example, only customers who have bought garden furniture will receive the sub-catalog. However, dynamic preferences are not taken into account when producing the sub-catalogs (Angermann, 2022). It looks identical for all recipients, regardless of whether a customer lives in an apartment with a balcony, or in a house with a garden and pool. ...
... For example, the product information on the e-commerce site is based on the same data as the content displayed in the printed product catalog. In most cases, the PIM and product taxonomy is used as the leading element to publish consistent product information across online and offline channels (Angermann, 2022). In terms of product master data, the external procurement relationship (EPR) is usually the leading system. ...
... The content-based analysis method uses natural language processing (NLP) to produce semantic correspondences of the original concepts in the form of mediator concepts. In TaxoCatalog, the output is based on the work presented in Angermann, Pervez and Ramzan (2017) and Angermann (2022). These are latent concept types of dependencies and collections (see Figure 4). ...
Taxonomies are a formal method of semantically structuring information using hierarchically ordered concepts. Those play a crucial role in omni-channel retailing to publish product catalogs across media channels, i.e. digital (portal based, paper-based) media channels, and print media (paper-based) media channels. Portal-based media channels use taxonomies to structure product related content by concepts to facilitate customers navigate through the e-commerce site. That are the product categories. Paper-based media channels require taxonomies for automatic layout setting using data-driven publishing software. When recommender systems are additionally used, products are published individually on the e-commerce site. Printed product catalogs, on the other hand, currently only display
content regardless of dynamic preferences, unless the layout is set manually. That is, as in an industrial context, personalization is of no interest if the necessary processes cannot be automated. The only recent industrially relevant method of taking preferences into account is to print sub-catalogs. However, these only contain certain product categories, resulting in a loss of information and sales, as preferences change dynamically today. With TaxoCatalog, an expert system is presented in this paper, capable of semantically personalizing product taxonomies to layout printed product catalogs according to the dynamic preferences of customers. The proposed expert system considers three layers to achieve full automation of relevant processes. The first layer uses background knowledge to consider a memory-based analysis of preferences, and a content-based analysis of possible semantic modifications. The second layer infers semantically personalized taxonomies using different modification rules. The third layer transforms the individual taxonomy paths into XML. This allows that the output of TaxoCatalog can be processed by any standard data-driven publishing software for automatic layout setting. A case study and a comprehensive evaluation, which discusses the strengths and limitations of previous research in the field, as well as the expert system in terms of quantitative and qualitative criteria, underline the efficiency of TaxoCatalog.
... In omni-channel retailing, the taxonomies are the most essential tool for publishing content across channels (Angermann, 2022). The data stored in the content management systems, is transferred via interfaces to the e-commerce portal. ...
Taxonomies are a formal method to semantically structure information using hierarchical ordered concepts. Those play a crucial role in omni-channel retailing to publish product catalogs across media channels, i.e. digital (portal-, paper-based) and print media (paper-based). Portal-based media use taxonomies to structure product-related content by concepts to facilitate customers navigating through the e-commerce site. I.e., the product categories. Paper-based media require taxonomies for automatic layout setting using data-driven publishing software. To achieve
consistency, all media use the identical product taxonomy as basis. When additionally using recommender systems, products are published individually on the e-commerce site. Paper-based
media, in contrast, show content regardless of dynamic preferences. Limited personalization takes place using sub-catalogs to only show products of, or specific categories. However, the resulted
information loss may result in sales loss, as preferences change. An approach to produce paper-based catalogs without information loss is not available. This is, as individual layout setting and printing is uninteresting, if automating the required processes cannot be achieved. With TaxoCatalog, an expert system environment is presented, which can automatically process customers’ preferences for producing individual product catalogs without information and
automation loss. Using TaxoCatalog, a semantic personalization is computed, before the layout is set. The first component of TaxoCatalog is semantically analyzing the initial product taxonomy
using natural language processing (NLP). The second component is analyzing customers’ preferences to create a personalized taxonomy based on the NLP outcome. The third component is used to scale the personalized taxonomy, depending on the desired volume of the catalog. The fourth component is used to perform the rule-based layout setting using standard data-driven publishing software. All decision processes are achieved using background knowledge to allow a cost-less integration into standard workflows. A comprehensive evaluation using two public and one private database provided by a German retailer underline TaxoCatalogs’ efficiency.
