Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Smart Shop Services for Building Customer-Oriented Scenarios

Krawczyk Henryk, Sobecki Andrzej

TASK Quarterly : scientific bulletin of Academic Computer Centre in Gdansk

2020

Vol. 24, No 3

221--242

The shops of today mostly support the customer by offering him or her products based on basic relationships between products viewed or ordered by users with similar tastes. This common approach may fail in many cases especially when the user does not have sufficient knowledge about the market, or when he or she wants to build a set of products in more than one shop. New categories of smart shop services are proposed in order to execute such customer-oriented scenarios where recommended products do meet mutual dependencies with products previously ordered by the customer. An attempt is made to collect additional information about the behavior of users (from past and current contexts) and represent it in a targeted graph called the customer-oriented scenario. Four types of such scenarios are distinguished depending on how many shops have been visited by the user before buying the expected products and how many products the user wants to buy. Moreover, the proposed scenario model provides the possibility of showing which services had been used by the user before the selection was made. Customer-oriented scenarios may be created post factum based on event data logs or before the user will use the shop, which means that it can be arranged which information, knowledge sources (internal or external), products or categories should be suggested in some context of the user's decision. The possibility of leveraging additional smart services into a traditional trading platform may help users, especially when they want to implement a complex scenario and order many products with mutual dependencies or in a situation when the user wants to understand the market before buying something. Using internal and external services allows creating a network for distributing knowledge focused on the actual customer context in a shop.

Towards semantic-rich word embeddings

Beringer Grzegorz, Jabłoński Mateusz, Januszewski Piotr, Sobecki Andrzej, Szymański Julian

Annals of Computer Science and Information Systems

2019

Vol. 18

273-–276

In recent years, word embeddings have been shown to improve the performance in NLP tasks such as syntactic parsing or sentiment analysis. While useful, they are problematic in representing ambiguous words with multiple meanings, since they keep a single representation for each word in the vocabulary. Constructing separate embeddings for meanings of ambiguous words could be useful for solving the Word Sense Disambiguation (WSD) task. In this work, we present how a word embeddings average- based method can be used to produce semantic-rich meaning embeddings, and how they can be improved with distance optimization techniques. We also open-source a WSD dataset that was created for the purpose of evaluating methods presented in this research.