Karrol
commented
3 months ago 我的代码如下
import Rbeast as rb # Import the Rbeast package as rb
from datetime import datetime
def cal_o(signal,tcp_max,scp_max):
ndvi_values = pd.Series(signal.values)
ndvi_index = pd.Series(signal.index)
调整图形大小
#plt.figure(figsize=(12, 8))
metadata = rb.args() ### or 'lambda: None': just get an empty object### # metadata is used to interpret the input data Y
metadata.time = 'hour'
metadata.season = 'harmonic' # fit a harmonic model to the periodic component
metadata.startTime = 1 # unknown unit
metadata.deltaTime = 1 # unknown unit
metadata.period = 24 # unknown unit [Guessed by autocorrelation]
metadata.maxMissingRate = 0.75 # if more than 75% of data is missing, BEAST will skip it.
metadata.deseasonalize = False # if true,remove a global seasonal cmpnt before running BEAST & add it back later
metadata.detrend = False # if true,remove a global trend cmpnt before running BEAST & add it back later版本1：变点数量搜寻范围
#o=rb.beast(ndvi_values[:],tcp_minmax=(0,tcp_max),scp_minmax=(0,scp_max),quiet=True, mcmc_seed=42)
#版本2：直接假设变点数量是多少
o=rb.beast(ndvi_values[:],tcp_minmax=(tcp_max,tcp_max),scp_minmax=(scp_max,scp_max),quiet=True, mcmc_seed=40) #随机数种子固定，这样可以对比不同的参数效果
return o def find_best_tcp_max(signal,search_tcp_max_range,search_scp_max_range): o_rmse_list=[] o_R2_list=[] for i in range(1,search_tcp_max_range): for j in range (1,search_scp_max_range): o=cal_o(signal,i,j) o_rmse=o["sig2"].RMSE o_R2=o["sig2"].R2 print(o["sig2"]. R2) print(o["sig2"]. RMSE) o_rmse_list.append([i,j,o_rmse]) o_R2_list.append([i,j,o_R2]) return o_rmse_list,o_R2_list
zhaokg
I try to find the best fitted model , so i use grid search based on diffrent scp_max and tcp_max combination. when I set parameters like :o=rb.beast(ndvi_values[:],tcp_minmax=(0,tcp_max),scp_minmax=(0,scp_max),quiet=True, mcmc_seed=42) ,it works well. But when i set parameters like :o=rb.beast(ndvi_values[:],tcp_minmax=(tcp_max,tcp_max),scp_minmax=(scp_max,scp_max),quiet=True, mcmc_seed=42), one of my time series will throw error r1<r2 during the circulate.
作者您好，我想请问一下这个问题为什么出现。我希望通过参数搜索找到rmse最低的拟合函数，然后我尝试了两种设置变点数量的方法。版本1是设置范围，版本2则直接制定变点数。也许是我搜索的范围过大而时间序列不够长吗？我在版本2的循环中获得了报错。