Grid Text Retrieval based on Deep Learning

The Internet of Things (IOT) have developed rapidly in recent years, which produces a vast amount of data including the electronic text data and the distributed data etc. These massive data hide a wealth of useful information and knowledge, and analyze them on the needs of users and assess the effect of products has a great significance. Therefore, it is an urgent problem to propose an efficient method of grid text retrieval. As is known to all, deep learning models are the most effective way to deal with massive data. In our paper, we will use an effective deep learning model to retrieve the grid text. i.e., the Deep Boltzmann Machines model (DBMs). The DBMs model is an effective text clustering algorithm. In particular, DBMs integrate a special energy function and use a layer-by-layer pre-training phase, which can make model more effective. The design of deep structure also enables the DBMs to extract the good text representations. The experimental results show that the DBMs model gets the desirable effect in handling with an extract abstract concept and has a good performance in large scale grid text clustering.