All Insights

Meet the Only Model That Automatically Recognizes Product Names from User Generated Content

Wissee AI-based Product Name Entity Recognition helps to identify hot products from millions of users’ social media postings. Our model beats the state-of-art named entity recognition (NER) methods.

09 Jun 2021 – Methodology

If you are wondering how to utilize the full power of user-generated content for making better decisions, read on. 

‘User generated content’ - or UGC, for short - refers to social media postings, blogs, reviews posted by consumers from online. It’s any and all content created and posted by users.

And UGC is powerful.

Nowadays, consumers’ purchasing decisions are strongly influenced by what they see on social media. This UGC impacts how people make decisions and spend money. So it’s no surprise that UGC data contains valuable information about consumer purchasing behavior.

But, because it’s generated by users all over the world, abbreviations, slang, internet jargon, typos, and non-standard syntactic constructions make it hard data to make sense of, let alone, make an investment decision. 

Wissee solves that problem so you can make sense of that data effectively and efficiently with powerful NLP and Machine Learning algorithms built as a result of over a decade of research.

Meet Wissee’s Product Named Entity Recognition.

It’s a Deep Learning AI model trained on over 100,000+ proprietary carefully human-annotated training data collected from multiple social media platforms.

And it is the only such model on the market that can automatically recognize product names from UGC text.

The Use of Product Name Recognition Model:

AI model for Product Name Recognition with Bi-LSTM and CRF using Word/Character Embedding

Wissee AI-based Product Name Entity Recognition helps to identify hot products from millions of users’ social media postings. Understanding consumer discussions on products can help investors and companies gain an early understanding of corporate financial performance and industry trends.

Our model beats the state-of-art named entity recognition (NER) methods

Stanford NER, IBM NER, and Wissee NER comparison