Zillow Taps AI to improve home value calculations
[ad_1]
Stories of people asking for the price of ten thousand dollars for money has become normal for their homes. This year, the U.S. housing market inventory has been a record high in house prices tall. Redfin CEO Glenn Kelman recently highlighted the madness tweet their first son offered to name the seller by telling the story of a home buyer — and they refused.
When the U.S. housing market began to heat up, in February Zillow began making initial cash offers based on home purchase prices. Now Zillow has updated its version algorithm behind these estimates, as the company says, will be more accurate and will allow Zillow to buy more homes.
Initially 900,000 homeowners were eligible to receive automatic cash offers to purchase a home. Stan Humphries, head of analytics at Zillow, said the change has changed Artificial intelligence this pool will increase by 30%. A company spokesman said Zillow Offers could close sales within a week.
Zillow previously determined the value of homes using nearly 1,000 variations of algorithms derived for local markets. Now all national prices will be decided by a single one neural network. Zillow said the new algorithm will reduce price estimation errors by 11.5% in market-free homes in nearly 30 U.S. regions. Compared to the previous version of the algorithm, the errors were reduced most in Phoenix, followed by San Antonio, Tampa, and Houston.
Using the new algorithm, Zillow will more frequently update estimates of the value of 104 million homes in 23 U.S. markets. The company was founded around 2005, where ratings were updated monthly. They have recently been updated several times a week; now some estimates can be updated daily.
Zillow’s house estimates are a well-known topic of local conversation, especially in the hot housing markets. A recently Saturday Night Live skit compared surfing to the sex of the Zillow phone and jokingly, for people in their thirties, “now the pleasure you get with sex comes from seeing other people’s homes.”
Zillow has been developing a home valuation algorithm for 15 years. The movement for a single neural network began in 2019 after a public competition to bring together more than 3,000 teams competing for the $ 1 million prize. Two of the three finalists in that competition used neural network-based approaches, Zillow says, which is an in-depth study that is able to know the relationship between the data used to calculate home valuation.
For example, Humphries said the new approach better understands the value of coastal assets or how valuations affect the size of nearby homes.
“The old approach would make it difficult to understand the value of coastal homes in that region,” he said. “Neural networks trained throughout the country can take information from other parts of the country about the value of the coast and apply it to local geography, even if there is no such home in that geography.”
House value calculations are based on tens of factors about a property, including square meters and location. Some lists contain data from tax assessments and sales records. Since 2016, the company has been using computer visual systems that draw conclusions from photos from home listings.
By expanding the home buying program, the company hopes to bring the one-click sensitivity of e-commerce to the real estate market, CEO Rich Barton said. “Ultimately, we want most homes in the country to have a direct offer to buy them inside that box,” he said.
The company also hopes to take advantage of what it calls a “big remodeling,” as more people are looking to work from home more often and millennials are buying homes.
Billions of people watch the Zillow list every year, many, Barton said, dreaming or performing a bit of voyeurism. For them, Zillow tries to “engage or entertain” them, but Zillow wants to sell more services to those who buy and sell houses. Beyond buying homes, Zillow helps homes with loans, home insurance and the fiduciary process.
More great KABEKO stories
[ad_2]
Source link