Here is something new from Gigaom: “Stanford Researchers To Open Source Model They Say Has Nailed Sentiment Analysis.” Richard Socher and a team from Stanford have created a computer program that can classify the sentiment of sentence with 85% accurately. They tested the model on movie reviews with a positive or negative tone. Even more amazing is that Socher and his team are making the project available to everyone. Why not capitalize on it instead? After all, companies have been trying for years to analyze social media and would pay the big bucks for said technology.
What makes Sucher’s project different from other sentiment software is that is reads whole sentences rather than just words.
“The team then built a new model it calls a Recursive Neural Tensor Network (it’s an evolution of existing models called Recursive Neural Networks), which is what actually processes all the words and phrases to create numeric representations for them and calculate how they interact with one another. When you’re dealing with text like movie reviews that contain linguistic intricacies, Socher explained, you need a model that can really understand how words play off each other to alter the meaning of sentences. The order in which they come, and what connects them, matters a lot.”
Socher hopes to reach a 95% accuracy, but the technology will never be 100% accurate because of jargon, idioms, odd word combinations, and slang. The project is making landmark strides in machine learning, logical reasoning, and grammatical analysis.
It means better news for online translators and speech technology, but commercial sentiment analytics vendors may see a decline in their profits.
Whitney Grace, October 21, 2013