A new algorithm called IDX has been developed to create better decision trees. Unlike existing methods, IDX focuses on improving decision tree quality by considering factors like tree size, error rate, and classification cost. By adjusting the complexity of the algorithm, decision trees of varying quality can be generated. It is important to address quality during tree construction as fixing mistakes later may not be effective. The algorithm's complexity, the underlying heuristic, and the results of experiments are discussed in the article.