Data classification is one of the most challenging areas in the field of Machine Learning and Pattern Recognition application where data is represented as a point in high-dimensional space. The data can be classified using supervised learning if it is already labeled. Otherwise unsupervised learning is used. To get golden point between them, Semi supervised learning is introduced which uses both labeled and unlabeled data. Analyzing the high dimensional data is the biggest challenge that can be tackled with the help of dimensionality reduction techniques. When Dimensionality Reduction is embedded in Semi supervised learning, it gives superior performance. The purpose of dimensionality reduction is to reduce complexity of input data without losing important details. In this paper, Semi supervised learning is studied using four different approaches. Analysis and comparative study of these techniques is illustrated with the help of three datasets. Role of dimensionality reduction is also observed in the classification of data.