Background/introduction: In recent years, stock market forecasting has received a lot of attention from researchers. This attention and the growing stock market investments have highlighted this as an important and emerging application of machine learning.Methods: In this research work, we present a stock trend forecasting system with a focus on reducing the amount of sparseness in the data collected using machine learning. We conduct an outlier detection of the data available for reducing dimensionality and implement a K-nearest neighbor algorithm to classify stock trends.Results and conclusions: The experimental results show the performance and effectiveness of the proposed trend forecasting system compared to the existing systems. The proposed system’s model (i.e., KNN classifier) gives better results of low error (MSE = 0.00005, MAE = 0.005 and Logcosh = 0.004) on KSE dataset as compared to previous works.