利用终端位置时空转移概率预测通讯基站服务用户规模
|
方志祥, 倪雅倩, 张韬, 冯明翔, 于冲
|
Using Terminal Location Spatio-temporal Transfer Probability to Predict Subscriber Base Size of Communication Base Station
|
FANG Zhixiang,NI Yaqian,ZHANG Tao,FENG Mingxiang,YU Chong
|
|
表2 3种模型的准确率对比 |
Tab. 2 Accuracy results of three prediction models |
|
时刻 | 时空转移概率模型 | | Castro的模型 | | 移动平均法 | 均值 | | 中值 | Q3-Q1 | 均值 | 中值 | Q3-Q1 | 均值 | 中值 | Q3-Q1 | 8时 | 0.950 | 0.956 | 0.064 | | 0.896 | 0.907 | 0.105 | | 0.819 | 0.821 | 0.171 | 9时 | 0.963 | 0.962 | 0.051 | 0.939 | 0.942 | 0.079 | 0.821 | 0.821 | 0.165 | 10时 | 0.966 | 0.965 | 0.051 | 0.943 | 0.944 | 0.079 | 0.800 | 0.805 | 0.179 | 11时 | 0.963 | 0.962 | 0.050 | 0.943 | 0.944 | 0.078 | 0.802 | 0.811 | 0.175 | 12时 | 0.955 | 0.955 | 0.058 | 0.937 | 0.940 | 0.100 | 0.778 | 0.791 | 0.197 | 13时 | 0.966 | 0.966 | 0.048 | 0.942 | 0.943 | 0.094 | 0.774 | 0.788 | 0.196 | 14时 | 0.964 | 0.964 | 0.054 | 0.941 | 0.941 | 0.084 | 0.777 | 0.788 | 0.192 | 15时 | 0.966 | 0.967 | 0.051 | 0.947 | 0.950 | 0.075 | 0.782 | 0.792 | 0.186 | 16时 | 0.961 | 0.964 | 0.054 | 0.941 | 0.943 | 0.078 | 0.791 | 0.800 | 0.182 | 17时 | 0.957 | 0.960 | 0.057 | 0.918 | 0.925 | 0.092 | 0.786 | 0.803 | 0.183 | 18时 | 0.946 | 0.948 | 0.079 | 0.915 | 0.920 | 0.131 | 0.802 | 0.799 | 0.180 | 19时 | 0.948 | 0.951 | 0.080 | 0.895 | 0.906 | 0.192 | 0.798 | 0.799 | 0.177 | 20时 | 0.965 | 0.963 | 0.050 | 0.922 | 0.916 | 0.128 | 0.785 | 0.790 | 0.181 | 21时 | 0.968 | 0.967 | 0.050 | 0.903 | 0.906 | 0.142 | 0.770 | 0.782 | 0.203 |
|
|
|