Drugi jezik na kojem je dostupan ovaj članak: Bosnian
Microsoft is using advanced tech to predict the results of this year’s Eurovision, the biggest singing contest in Europe. Using predictive algorithms in real time, based on the factors and circumstances which are continually changing, Microsoft found that Croatian representative Jacques Houdek with the song My Friend has a 0.7% chance for the overall victory which together with the Swiss, Danish and Armenian representatives places him at 13 place. The greatest chances for overall victory has the Italian representative Francesco Gabbani with the song Occidentali’s Karma, and the chances to win of 18%. He is followed by the Swedish representative Robin Bengtsson with 16.8% and the Bulgarian Kristian Kostov whose chances are 10.2%. Other candidates to win in the top ten are all Belgium, Portugal, Montenegro, Azerbaijan, France, Albania and Finland, in that order.
Prediction is based on three key analyses – data from the Internet, data from social networks and search traffic data. In this process it is important to conduct an analysis in real time because, for example, various signals from the Internet and social networks during the event are correlated with the votes of the audience. Microsoft uses these signals and compares them with the relevant data from the past, as for example which country usually votes for which country, and thus determines different patterns and predicts the most likely results. Otherwise, when it comes to a variety of events, Microsoft’s prediction rate is extremely high. Example are recent successful results for the Formula 1 race in Abu Dhabi in 2016, where Microsoft successfully predicted all the drivers who finished in the top five positions, or the Academy Awards where the rate of predicting the winners by category were as high as 71%.
Algorithms that Microsoft uses for forecasting play an important role in various other industries that are currently going through a phase of digital transformation. The same algorithms can be used by companies of all sizes, and the data provides invaluable insight into what people are searching on the internet and what they are thinking. Algorithms for example also provide help to companies to determine the best time to launch a new product and gain insight into potential problems related to the product.