By using a series of Neural, Sentimental and Linguistic cues, Biasly´s proprietary algorithm can analyze the components that make up a certain political leaning either liberal or conservative, in real time. The Artificial Intelligence powered software, does this by comparing certain aspects of political policies and political entities such as politicians, as well as sentiment with the parameters expressed in an article to then produce a score between -100% or 100%. Scores of -100% to 0 mean that the article leans more towards the liberal side of the scale, while a rating of 0 to 100% means that said article/author/website/politician is more conservative.
The Bias Meter, analyzes articles to be able to determine which policies are being discussed, which politicians are mentioned, author bias leaning and different priority aspects for certain political parties, to then determine the overall sentiment expressed in an article to then give the article its due rating.
Our proprietary algorithm was developed by data science experts who have specialized in predictive analysis, Python convolutional neural networks, among many others. These specializations are used to enhance machine learning and better predict biases in the media. As data enters and passes through the connected networks, the data is then transformed into a score on the meter. At Biasly, we use natural language processing (NLP) and entity specific sentiment analysis to transform said data. The aforementioned tools can then analyze text to see how the positive or negative sentiment affects the text. This process is done by the use of algorithms that learn a desired output, by analyzing a large amount of data that functions as an example of the desired output.