移动大数据不仅具有“4V”特性，还表现出明显的个性化、时空性和层次性，给现有的数据分析理论带来了严峻挑战。本项目以深度融合人机能力为核心，以需求牵引的多粒度分析为主线，研究基于人类认知机理的移动大数据高效分析方法。主要研究内容：(1)基于人类认知机理的类人认知计算模型，使计算机能模拟人脑思维方式来分析移动大数据；并通过自主学习，不断提升其分析能力。(2)需求自适应的多粒度分析模型，借鉴人类层次化分析问题的机理，研究需求驱动的多粒度移动大数据分析框架。(3)人机深度融合机理，研究如何充分融合人类非凡的感知、认知能力与计算机强大处理能力。(4)基于人机认知融合的多粒度移动大数据分析方法，研究深度自适应机制以及精确与非精确计算相结合的方法，实现用户兴趣及其时空特征的高效挖掘。本项目将建立基于人机融合的多粒度移动大数据分析若干理论与关键技术，形成创新的移动大数据分析理论成果。

Mobile big data exhibits personality, space-time and hierarchy features besides the traditional “4V”, which brings serious challenges to existing data analysis theories. Focusing on deeply fusing abilities of human and computers and demand-driven multi-granularity analysis, this project will explore efficient analysis methodology for mobile big data based on human cognition mechanisms...In this project, we will intensively investigate the following important issues.Firstly, we are going to propose a human-like cognition computing model based on human cognition mechanisms. It enables computers to understand mobile big data in the human cognition way, and increasingly improves analysis ability through continuous self-learning. Secondly, we will set up a demand-adaptive multi-granularity analysis model. A multi-granularity analysis architecture for mobile big data will be proposed, based on the human multi-layer cognition mechanism. Thirdly, we will explore deep human-computer fusion mechanisms to make full use of excellent sensing and cognition capabilities of human and powerful processing capability of computers. Finally, a multi-granularity analysis methodology for mobile big data will be set up. For this purpose, we are going to investigate depth-adaptive mechanisms, as well as methods combining precise and imprecise computing. Consequently, a set of efficient algorithms for mining interests and space-time properties of mobile users will be developed...This project will found a set of new key analysis technologies for mobile big data based on human-computer fusion, and will deliver innovative academic results for mobile big data analysis.