Our presentation at Text Mining & Application Workshop in Ho Chi Minh City!
Real Estate has become the hot market in the modern society. People commonly search for suitable accomodation through traditional channels such as the Internet and realtors. Most of the many websites dedicated to trading real estate require users to specify their desired criteria in a predefined form. However, natural languages are preferred by most people when advertising or searching for a real estate property. With that said, real life real estate search as well as advertising are mostly unstructured and naturally ambiguous. This poses a real challenge for automatic processing.
In our #AcademicPaper, we consider this problem as a sequence labelling task and apply a Deep Learning model to extract features of a property (i.e. address, type of property, etc.). We adopt the current state-of-the-art Deep Learning architecture, which includes word-level and character-level Convolutional Neural Networks, bidirectional Gated Recurrent Unit, Conditional Random Field and Word Embedding. The advantages of our approach have been proven when experimented with real data collected from real estate trading sites.
Don’t miss out. Take a peek at our AcademicPaper now.
Have a question?
Don’t hesitate to join our Telegram and ask our friendly admins, @linachloedo and @HungBlue.
Let’s Activate the Power of Business Network!
BitCEO + Zeniius Team
#Tokensale #ArtificialIntelligence #Blockchain #Cryptocurrency #Smartcontract #CEO Network
