Newer perspectives toward technology aim to leverage current methods to achieve even greater heights. Machine learning, a game-changer in the real world, assists machines in progressively improving their performance. It also enables software applications to accurately predict outcomes without explicit programming. Traditionally, machine learning algorithms use past data for predictions. In a typical machine learning environment, the primary focus is on data training. The collected data is divided into two purposes: one for training and the other for testing. Common practice suggests reserving a higher percentage of data for training the machine because the more the machine is trained, the better its results become. While the machine learns from the training data, its capabilities are assessed based on its performance on unknown data. If the machine can produce convincing results on unknown datasets, it is likely to perform well on real-time data. As a result, data testing takes centre stage. Therefore, in this paper, the proposed idea is of using limited yet efficient datasets for training and reserving the rest of the data for testing the machine.