Feature selection is an important process in machine learning, whose goal is to find a small number of features by removing irrelevant and/or redundant features. The main idea of feature selection is to choose a subset of input variables by eliminating features with or no predictive information. Feature selection removes dimensionality reduction problem. It is best for high dimensional data. Feature selection methods can be divided into three classes. One of is filter methods and another one is wrapper method and the third one is embedded method. This is one of two ways of avoiding the curse of dimensionality. This work reviews several fundamental feature selection algorithms found in the literature.