In this era of data where choices are way more than the number of users associated with it, the role of recommendation and ranking dominates one's selection. As with every spare second spent over the network, data generated by the interaction increases so rapidly. Through this data it becomes a task of just applying advanced algorithms to generate propaganda oriented campaigns for the welfare of any particular organisation associated. These organisations use this to manipulate the selections. With this paper we aim to develop a minimal error recommendation and ranking system that provides unbiased and desired results with the help of integrated machine learning techniques and neural network interconnections. We are considering big-data sets that are cloud based and oriented to pre-filtration and generalise categorisation of fetched data from the user, which contributes in speeding up the learning process of the module.