From: Providing privacy preserving in next POI recommendation for Mobile edge computing

N | number of user states in the model |
---|---|

q_{i} | The i_{th} user’s state |

Q_{u} | Q_{u} = {q_{1}, q_{2}, …q_{N}}, the state set of user u |

M | number of POIs visited by user. |

t_{i, j} | the probability that the state q_{i} transforms to the state q_{j} |

T | T = {t_{i, j}}_{N ∗ N}, state transition probability distribution |

L | length of the POI sequence |

O | O = {I_{1}, I_{2}, …, I_{M}}, the POI sequential data |

p_{i, j} | the probability that user visit POI I_{j} when the state is q_{i} |

P | p = {p_{i, j}}_{N ∗ M}, preference probability distribution |

π | the probability of initial state |