University of Twente Student Theses
Twitter- a new pathway to access product innovation ideas - Can machine learning help PepsiCo identify innovative ideas in User-Generated Content platforms?
Chen, Yucheng (2022) Twitter- a new pathway to access product innovation ideas - Can machine learning help PepsiCo identify innovative ideas in User-Generated Content platforms?
|
PDF
812kB |
| Abstract: | In the last decades, fulfilling ever-changing customer needs in this changing market environment has been chased by multiple commercial entities, which has created enormous opportunities for machine learning-based User-Generated Content (UGC) analysis. For information extraction tasks from unstructured UGC, natural language processing (NLP) approaches based on machine learning are becoming increasingly popular for information extraction tasks from unstructured UGC. The objective of this paper is to quantify and enhance Text CNN’s text classification performance on tweet datasets generated by the Twitter keywords filter. It will discuss the primary performance of Text CNN and the problem this model faces during dealing with the task of identifying numerous tweets. Through discussing the different performance metrics, it suggests using a macro F1 score as a baseline and concentrating on improving the recall of the class “without CNs”. Then by identifying problems like imbalanced datasets, this paper discusses previous solutions. Furthermore, this research also intends to analyse the model’s application to PepsiCo’s flavour innovation processes through the extracted information. It also suggests that PepsiCo may empower the cost-efficiency of new flavours market research via adopting the machine learning-based UGC analysis. Lastly, recommendations based on observations will be provided for further research on the application of machine learning models in the beverage industry. |
| Item Type: | Essay (Bachelor) |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 85 business administration, organizational science |
| Programme: | International Business Administration BSc (50952) |
| Link to this item: | https://purl.utwente.nl/essays/91018 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page