OREANDA-NEWS.  Yandex today launches its new machine learning-based technology for hyper-local and super-accurate weather forecasting. The technology, called Meteum, collects a massive volume of data about past forecasts made by external providers like WRF or Foreca, and compares them with weather conditions of the past to figure out the difference between what was forecasted and what happened in reality. It then compiles a mathematical formula to correct and refine future forecasts. Here is a diagram of how it works.

Meteum is now the core technology for the updated Yandex.Weather web service and mobile app available for iOS and Android. Meteum calculates a new forecast each time a user consults Yandex.Weather. It locates a person and shows them a fresh forecast for precisely that spot. The user can choose another place and time for the forecast to see what the weather will be like around their office in an hour or whether it might rain when they go out of town in the evening.

More details on how Meteum works can be found in Yandex’s official blog.