An Improved Deep Learning Model for Word Embeddings Based Clustering for Large Text Datasets