With the advent of automatic learning methods and the exponential growth of computer power, several things are being facilitated, in particular the prediction of certain behaviors, made by banking establishments. Although some prediction tools such as Excel are nowadays used by some of our establishments, we find that machine learning remains unused so far, even less with all its power, yet many advantages and opportunities present themselves to it. use. We argue that elsewhere, experiments based on machine learning, in other words automatic learning, are more than topical, even more so in the banking sector. We are therefore going through this experience, to propose as an illustration and educational, for our local banking establishments, an activity of prediction of monetary inflation, using the power of artificial intelligence. The prediction in question here will be made on the basis of a set of data collected from a few banks in the city-province of Kinshasa, and particular emphasis will be placed on general consumer price indices. It is important to note that the prediction in question here goes as far as clearly specifying the causes of the inflation being analyzed, or predicted, of course on the basis of the different variations of the indicators. The machine learning used here offers us several possibilities in terms of algorithms and models, but in the context of this work, we will only address a few, in particular Linear Regression, the Random Forest Regression algorithm or Radom Forest Regression, and the Regression model decision tree, will get we will get the best algorithm with respect to its score.

The idea of linear coding is simple: by an injective linear application we send a space of binary words in a larger space, hoping that the redundancy introduced helps us to detect and correct the transmission errors. Among the so-called linear codes, we consider in the context of this article the Hamming code which is a perfect code, because for a given code length there is no other more compact code having the same capacity correction. In this sense its yield is maximum. In this work, we have proposed an algorithm based on the above characteristics of the Hamming code, which we can implement in a given programming language.