Abstract:Statistical models based on Markov chains, especially mixtures of Markov chains, have recently been studied and demonstrated to be effective in various data mining applications such as tourist flow analysis, animal migration modeling, and transportation administration. Nevertheless, the research so far has mainly focused on analyzing data at individual levels. Due to security and privacy reasons, however, the observations in practice usually consist of coarse-grained statistics of individual data, a.k.a. aggregate data, rendering learning mixtures of Markov chains an even more challenging problem. In this work, we show that this challenging problem, although intractable in its original form, can be solved approximately by posing structural constraints on the transition matrices. The proposed structural constraints include specifying active state sets corresponding to the chains and adding a pairwise sparse regularization term on transition matrices. Based on these two structural constraints, we propose a constrained least-squares method to learn mixtures of Markov chains. We further develop a novel iterative algorithm that decomposes the overall problem into a set of convex subproblems and solves each subproblem efficiently, making it possible to effectively learn mixtures of Markov chains from aggregate data. We propose a framework for generating synthetic data and analyze the complexity of our algorithm. Additionally, the empirical results of the convergence and the robustness of our algorithm are also presented. These results demonstrate the effectiveness and efficiency of the proposed algorithm, comparing with traditional methods. Experimental results on real-world data sets further validate that our algorithm can be used to solve practical problems.

State aggregations in Markov chains and block models of networks

Markov Chain Aggregation for Agent-Based Models

Dynamical mixture modeling with fast, automatic determination of Markov chains

Information-Preserving Markov Aggregation

Adaptive Low-Nonnegative-Rank Approximation for State Aggregation of Markov Chains

Formal Error Bounds for the State Space Reduction of Markov Chains

A Dynamic Stochastic Block Model for Multi-Layer Networks

Empirical Bayes estimation for the stochastic blockmodel

Stochastic Block Models are a Discrete Surface Tension

Agent based models and opinion dynamics as Markov chains

Learning Mixtures of Markov Chains from Aggregate Data with Structural Constraints

Maps of sparse Markov chains efficiently reveal community structure in network flows with memory

Multilevel Adaptive Aggregation for Markov Chains, with Application to Web Ranking

Longitudinal Network Models and Permutation-Uniform Markov Chains

Aggregative Efficiency of Bayesian Learning in Networks

Large Deviations Properties of Maximum Entropy Markov Chains from Spike Trains

Mutual information for the sparse stochastic block model

Temporal aggregation of Markov switching models

Computational and Statistical Thresholds in Multi-layer Stochastic Block Models

Dynamic stochastic blockmodels: Statistical models for time-evolving networks

Maximum likelihood degree of the $β$-stochastic blockmodel