In the United States, those who suffer from seasonal allergies have one more tool to better manage their day-to-day lives, this time thanks to artificial intelligence (AI). The initiative of the team of IBM Watson and The Weather Channel, which launched a new feature in the app and on the website of the weather channel, providing information and alerts with a 15-day forecast.
In the middle of spring and with the summer approaching, the television channel's data team has been working to create a resource for those suffering from seasonal allergies, explains IBM in a blog post. In the investigation, experts conclude that about 60% of the people surveyed assess the risk of contracting allergy mainly based on three aspects: presence of trees, a very popular plant in the United States and grass.
However, The Weather Channel team ensures that plenum sources are not "reliable" and that they only cover a small subset of species that can lead to allergic reactions. The conclusion is that the plenum "is not a good indicator for assessing the risk of seasonal allergies or how the person feels" and this is where AI plays a key role.
In the study, the team concluded that the use of AI and meteorological data, instead of just using information on plenum levels, performed better. According to the scientists, this method has resulted in an increase of 25% to 50% in making better decisions regarding allergy risks, such as taking medications to prevent allergies.
IBM Watson's machine learning allows training of the local data model
The team used the machine learning of the IBM Watson supercomputer to train the local model of the data, in order to "help provide a forecast capable of assessing the underlying conditions that cause allergy symptoms".
The functionality now launched combines information from the TV channel, from temperature, humidity levels, precipitation and wind, with health and location data from IBM. The company guarantees that this information is anonymous, and is used only to better understand the environment and the local flora. The data on the plenum and air quality were excluded from the predictive model, since they "proved to be unreliable indicators in risk assessment. seasonal allergies, but will continue to be reviewed ", guarantees IBM.