:laughing:

We attribute these discrepancies to the limited number of stations in the area. the computational cost of the UKI is primarily dictated by the size of the model ensembles associated with the sigma points multiplied by the number of iterations. With 19 model parameters, UKI conducts 39 model evaluations in each iteration, and we run 10 iterations to obtain the final result, thus processing 380 model evaluations in each prior setting. Through extensive testing, it demonstrates that UKI requires minimal prior information about S-wave velocity for joint inversion of RF and SWD, although some prior means regarding layer thickness are necessary. Conversely, the trans-dimensional methods excel in this aspect, they do not require specific layer counts but allows for a general range. Where \mu_n represent mean and covairan of Gaussian distribution pn in the nth iteration, \mu_n+1,C_n+1 denote the updates in the n+1th iteration.

This article requires a password to view.

发微博好像不太合适，来写博客更好一些。 2023再见😁 2023年吃不上肉本家了... 吱猪的2024年第一顿，浙大走起！

放个电视看一看

最近几天总算把服务器折腾好了，来庆祝一下伊立扎德奶站的重新开张🎇🎇🎇 附上最近的学生年会报告😁😁😁

plot 线型 linestyle '-'/':'/'--'/'-.'/' ' 共用坐标轴 axd["B"].set_ylim(axd["A"].get_ylim()) 自己设颜色 cmap = plt.cm.RdYlBu # define the colormap ## extract all colors from the .jet map cmaplist = [cmap(i) for i in range(cmap.N)] cn = cmap.N nhv = len(HVV) for i in range(nhv): abc=ax.plot(thv_series,HVV[i],c=cmaplist[cn//nhv*i]) 整体调整字体 parameters = {'axes.labelsize': 10, 'axes.titlesize': 35} plt.rcParams.update(parameters) 图片保存 bbox_inches=’tight’ 对坐标轴的细更改 from matplotlib.ticker import FormatStrFormatter 1. starte, ende = axd["E"].get_ylim() axd["E"].yaxis.set_ticks(np.linspace(starte, ende,6) axd["E"].yaxis.set_major_formatter(FormatStrFormatter('%.e')) 2. axd["F"].set_xlim(0,22) starte, ende = axd["F"].get_xlim() axd["F"].xaxis.set_ticks(np.arange(starte, ende,2)) axd["F"].xaxis.set_major_formatter(FormatStrFormatter('%d')) 调整坐标轴之间的间距 fig.tight_layout(w_pad=3) plt 初始化 parameters = {'axes.labelsize': 35, 标号刻度 'axes.titlesize': 35, 'xtick.labelsize': 20, x轴刻度size 'ytick.labelsize': 20, y轴刻度size 'font.sans-serif': "Arial", 'legend.fontsize': 16, 'legend.handlelength': 1.1, 'legend.shadow':0, 'legend.edgecolor':'k', 'legend.framealpha':0.90} 手动更改坐标轴 ！！！ 重要 ll, bb, ww, hh = axd["H"].get_position().bounds axd["H"].set_position([ll- 2 * ww, bb, ww, hh]) 颠倒图例顺序 handles, labels = axd["E"].get_legend_handles_labels() axd["E"].legend(handles[::-1],labels[::-1],fontsize=15,loc=4) 去除最后一个图例 han, lab = ax.get_legend_handles_labels() ax.legend(han[:-1], lab[:-1], loc=3 删除一个图 axd["I"].remove() 设置刻度为平均间隔 ## 设置刻度间隔 x_major_locator=MultipleLocator(30) ## 设置刻度 # axd["A"].xaxis.set_major_locator(x_major_locator) 设置确定的刻度(如台站距离) axd["A"].set_xticks(Dist,minor=False); axd["A"].set_xticklabels(stts,fontsize=10,rotation=270); 设置刻度字体大小 axd["C"].tick_params(labelsize=ftt) #刻度字体大小20 设置 log 刻度 axd["F"].set_yscale('log') 刻度是整数 from matplotlib.ticker import MaxNLocator plt.gca().xaxis.set_major_locator(MaxNLocator(integer=True)) plt.gca().yaxis.set_major_locator(MaxNLocator(integer=True))

插入域中的数字 插入->文档部件->[类别：编号]->[域名:Seq]->[域代码]:"SEQ"->"SEQ list" -> 复制 -> 更新域 也可以用代码操作如下 1. 切换域代码 Ctrl + F9 2. 加入代码({SEQ list \* MERGEFORMAT}) 3. 回复 Alt+Shift+u 加入章节号 给文章分节 布局-> 分节符 -> 下一页 插入 -> 文档部件 -> [类别：Section] -> [格式：...] -> 复制 -> 更新域 也可以用代码操作如下 1. 切换域代码 2. 加入代码({SECTION \* MERGEFORMAT} - {SEQ list \* MERGEFORMAT}) 3.对章节号进行操作 (={ SECTION }-1 \* MERGEFORMAT} - {SEQ list \* MERGEFORMAT}) 另外一些操作 对每个章节后的公式重新计数， 加入\r 更新域 F9 进入域 Shift+F9 使用短的- 如果你已经在公式编辑器中，但减号看起来仍然太长，你可以尝试以下方法： 选中减号。 然后按下键盘上的 Ctrl + Shift + -（减号键）来插入一个短的减号 Reference [1]https://www.bilibili.com/video/BV16B4y1S7Ky?share_source=copy_web [2]https://bettersolutions.com/word/fields/seq-field.htm [3]https://www.bilibili.com/video/BV1Rb4y1Z7WC/?vd_source=692865acdc6cb0c0e206b2f487199950

Matrix mulitply python a1 = np.arange(2) # [0,1] (2,) a2 = np.arange(6).reshape(2,3) # [[1,2,3][4,5,6]] (2,3) a1@a2 > array([3, 4, 5]) # (3,) a2[None,:]@b > array([[3, 4, 5]]) # (1,3) a1 * a2 >array([[0, 3], [0, 4], [0, 5]]) Q: what's the order of a a2 @ a2.T: a1 a2 @ a2.T = (a1 * a2 ) @ a2.T 实质，将系数a1平均分给a2的列向量上 Julia CPP Cholesky Decompostion The cholesky decompostion is exclusively defined for symmetric or Hermitian positive definite matrices, A = LL*. Python: 1. chol_xx_cov = np.linalg.cholesky(x_cov) 2. s1,v1,d1 = np.linalg.svd(x_cov) v1 = np.clip(v1,a_min = 1e-8,a_max = None) q1 = np.linalg.qr(np.sqrt(v1)[:,None] * d1)[-1] chol_xx_cov = q1.T Julia: CPP

Conda This is a summary about conda/jupyter commonly used commands. Installation miniconda https://mirrors.tuna.tsinghua.edu.cn/help/anaconda conda initialize conda init Create environment and packages conda create -n env_name list of packages conda create -n py python=3.3 conda create -n data python=3.5 numpy pandas For package conda install/upgrade/remove/search package_name conda upgrade --all # update all package For environment view the environment conda info --envs activate environment Linux: source/conda activate my_env Windows: active my_env deactivate environment Linux: source deactivate my_env Windows: deactivate my_env delete environment conda remove -n your_env_name --all list package conda list delete environment conda clean --packages --tarballs kernel list kernel jupyter kernelspec list delete kernel jupyter kernelspec uninstall python37564bitbasecondac521c0e867d74695b0ab483942c57dc3 add environment conda install nb_conda or pip install ipykernel python -m ipykernel install --user --name pygmt --display-name pygmt change channel channels: - https://mirrors.ustc.edu.cn/anaconda/pkgs/main/ - https://mirrors.ustc.edu.cn/anaconda/cloud/conda-forge/ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/ - defaults show_channel_urls: true conda info change font conda install -n env-name -c conda-forge mscorefonts