Explore an introductory overview of efficiently representing text and images for predictive analytics in this 29-minute conference talk by Steve Gallant at BDF 2015. Discover techniques for transforming text of varying lengths and document structures into single, distributed, fixed-length vectors, enabling machine learning approaches for predictive modeling. Learn about applications such as sentiment analysis, news story and email classification, marketing models, and anomaly detection. Witness a live demonstration of classifying news stories over the web. Gain insights from Dr. Gallant's extensive experience in neural network learning, text representations, and machine learning, supported by his work at MultiModel Research and a National Science Foundation grant.
Overview
Syllabus
Introduction
Marketing
Language
Representation
Neural Motivation
Computational Motivation
Similar Vectors
Document Retrieval
Structure
Representation Approach
Vector Quantization
Deep Learning
Positional
Taught by
Open Data Science
