%0 Conference Proceedings
%T Package and Classify Wireless Product Features to Their Sales Items and Categories Automatically
%+ Nokia [Finland]
%A Tang, Haitao
%A Eratuuli, Pauliina
%< avec comité de lecture
%( Lecture Notes in Computer Science
%B 3rd International Cross-Domain Conference for Machine Learning and Knowledge Extraction (CD-MAKE)
%C Canterbury, United Kingdom
%Y Andreas Holzinger
%Y Peter Kieseberg
%Y A Min Tjoa
%Y Edgar Weippl
%I Springer International Publishing
%3 Machine Learning and Knowledge Extraction
%V LNCS-11713
%P 317-332
%8 2019-08-26
%D 2019
%R 10.1007/978-3-030-29726-8_20
%K Natural Language Processing
%K NLP
%K Machine Learning
%K ML
%K Process automation
%K ML based decision making
%K LTE
%K 5G
%K Business Digitalization
%K Pricing
%Z Computer Science [cs]Conference papers
%X Aiming at automated decision making, this paper defines and analyzes two machine learning use cases for the product process in wireless infrastructure business. The first use case assigns a product to a product packet according to the functionality of the product. The second use case determines the category of the product so that it can be priced. Then, the product is ready for sale. This paper also provides solutions to these machine learning use cases. The solutions are examined with real data from the processes. The credibility of the solutions is also evaluated by comparing the machine learning decisions with the decisions of human users. These human users know the actual assignment and classification of those products. The results show that the solutions work well as they expected. These solutions assign and classify a part of the given products fully automatically with a high confidence and accuracy. Due to insufficient prediction confidences for the rest of the given products, the rest part of products needs to be escalated for the further decision by the human users. With an escalation, a set of assignment and classification options for a given product is also recommended by the solutions. Often, the correct assignment and classification exist in the set of options already. The human users can easily identify and select the correct assignment and classification from the recommended options. Significant costs and processing time can thus be prevented.
%G English
%Z TC 5
%Z TC 12
%Z WG 8.4
%Z WG 8.9
%Z WG 12.9
%2 https://inria.hal.science/hal-02520039/document
%2 https://inria.hal.science/hal-02520039/file/485369_1_En_20_Chapter.pdf
%L hal-02520039
%U https://inria.hal.science/hal-02520039
%~ IFIP-LNCS
%~ IFIP
%~ IFIP-TC
%~ IFIP-TC5
%~ IFIP-WG
%~ IFIP-TC12
%~ IFIP-WG8-4
%~ IFIP-WG8-9
%~ IFIP-CD-MAKE
%~ IFIP-WG12-9
%~ IFIP-LNCS-11713