Loss Func APK

In the field of economic and administrative sciencies, decisions are always surrounded by uncertainty factors. In any decision under conditions of uncertainty, the first step to construct an analytical scheme that allows to delimit the possible economic losses is to define what is known as the loss function; in the decision, the best possible option will always depend on the way in which the different types of errors that an individual can commit are weighted; that is why the construction of the loss function is the ideal resource for this purpose.
In general terms, it can be said that in the loss function there are two basic arguments: the true class and the predicted class; both can be represented in a matrix.

L(y,y_1 )
Where L is the loss function, y is the true class and y1 is the predicted class.
If one thinks, for example, of undertaking in the field of technological innovation and in doing so happens that there is not enough market for the new products, this wrong decision will have a cost; however, if there is indeed a market for the new product and the decision of undertaking had not been taken - despite that there were indicators about that it was advisable - then the entrepreneur would realize that he lost millions because he did not want to launch their products at the right time , while the competition did.
This type of reasoning can be used to weigh different errors and guide decision making. Because it is the simplest, a binary loss function is normally used for such purposes; that is, a function that can assume only the values of 0 and 1. Thus, the decision can be made in the following terms: nothing is added to the cost if the correct decision is made and the unit is added in case of having made the wrong decision. This app facilitates the construction of loss functions for all kinds of decision problems.