How It Works

Aspectiva has developed unique and advanced NLP and Machine Learning technologies that analyze customer opinions from across the web, automatically identifying the topics being discussed, as well as understanding the sentiment expressed about each individual topic. Then, Aspectiva’s engine automatically generates scores for each product aspect, resulting in an at-a-glance crowd opinion overview of the product. Additionally, Aspectiva suggests the most relevant products that best fit shoppers’ individual needs.

Challenges

Aspectiva’s state-of-the-art Natural Language Processing algorithms robustly process unstructured data in any domain and automatically provide accurate, multidimensional, high-resolution analyses leading to actionable insights.

Examples of Challenges we Meet

UNDERSTANDING IF SENTIMENT IS POSITIVE OR NEGATIVE ACCORDING TO CONTEXT
“low quality” vs. “low price”
UNDERSTANDING SENTIMENT FROM BASIC FACTUAL DESCRIPTIONS
“The headset falls out of my ears.” “The Roomba cleans under the bed.”
DISTINGUISHING SENTIMENT FROM CONTEXTUAL DESCRIPTIONS
“It’s great for difficult corners.” “I have terrible allergies, and this cream works like magic.”
UNDERSTANDING NEGATION
“How can you not enjoy the beautiful sea”
“I don’t think Windows 8 plays nice with this hardware”

Technology Highlights