To deal with a huge amount of unstructured text data, PERCENT collects and analyzes public opinion information and communicationtrend in media and communities by its ability to process a huge amount of data in order to manage public opinions more effectively.

It uses the advanced deep learning technology to get the positive and negative sentiment value of specific text in different fields. Then the workshop adopts its advantages and avoids its shortcomings to realize continuous upgrades of product and service qualities.

It keeps trace of media and public opinion hotspots through automatic text summary and hotspot evaluation technologies to grasp critical content, efficiently extract information, and accurately explore text value.

It builds suspects’ communication network graph through big data in tele-communication, such as in mobile phone, QQ, WeChat, text message and so on. Besides, the workshop can find out the key personnel of the gang through SNA, complex network and related knowledge of graph theory，such as backbone members, important contact persons and so on.

By using machine learning algorithm, such as classification, clustering and regression, the workshop can extract critical variables through feature engineering and cross compare the case in progress and previous cases to find out whether the suspect has criminal record and provide related clues.

By integrating, processing, analyzing and modeling for a huge amount of multi-source public security data and government data, the workshop can make real-time warning when abnormal gathering occurs within a certain time in a certain area and realize crime forecast.

In the industrial big data era, enterprises face problems of complex system maintenance and guarantee. Combining the international cutting-edge feature engineering, hypersphere model (the company has applied for patent), and prediction algorithm model of deep neural network, the device data and business are deeply integrated. The workshop can accurately predict the operating status of the equipment and its components through PHM technology (equipment fault diagnosis, prediction and health management), and make real time warning so as to reduce equipment repair cost significantly and improve repairing accuracy. 

PERCENT pays attention to bank, insurance and security industry, collects multidimensional high-quality data features and extends the range of customers’ information. Besides, effectively combined with financial business experiences, PERCENT can explore the value of the data to the most, depict the customers’ truest images and features, deeply search for the demands and lifecycle of current and potential customers, and comprehensively understand the core attributions and association of the product. The workshop can improve marketing efficiency and accuracy by data mining and consulting technologies, and help traditional finance enterprises manage users and optimize products scientifically and reasonably.