Page 7 - summer of ai
P. 7

 is a type of AI that imitates domain experts. Specifically, expert systems have a knowledge base that serves as a proxy to the factual domain knowledge possessed by the human experts and an inference engine that emulates their analytical thinking. Several algorithms were built to create knowledge bases. For example, N-Grams is one of the algorithms that parses text or speech, and learns patterns among the phrases in the data. The patterns learned from the perfect data are used to standardize imperfect data.
Results from the latest lost sales analysis study found 6,791 standardized lender names cover
75% of DMV records (a drop from the initial 229,219 raw lender names). After looking at over 170 million DMV records, the team found the expert system reduced the number of lender names with only 1-2 DMV records by 54%.
AI that learns from experts improves the quality of the data during the data ingestion phase.
The improved data enables business analysts to more accurately answer questions such as: If a client didn’t buy a car from me, where did the client buy the car? Which lender originated the auto loan? As a starting point, Dr. Raj, Piyush and others in the Data Science Lab use NLP to perfect
TradeSight
Introducing a New Dealer Market Intelligence Platform
7
       ingested
data so our customers can make smarter decisions in the second phase of the insights supply chain: Insights Generation. g
DR. RAJKUMAR BONDUGULA PRINCIPAL DATA SCIENTIST
Download Product Sheet
With detailed information on more than 67 million auto tradelines
  





















































































   5   6   7   8   9