Author: bugman
Date: Thu Oct 2 13:08:52 2014
New Revision: 26131
URL: http://svn.gna.org/viewcvs/relax?rev=26131&view=rev
Log:
Merged revisions 26095 via svnmerge from
svn+ssh://bugman@xxxxxxxxxxx/svn/relax/trunk
........
r26095 | tlinnet | 2014-09-30 10:41:07 +0200 (Tue, 30 Sep 2014) | 5 lines
Implemented counting of outliers for statistics.
This is to get a better feeling why some statistics are very much different
between NI.
Task #7826 (https://gna.org/task/index.php?7826): Write an python class for
the repeated analysis of dispersion data.
........
Modified:
branches/frame_order_cleanup/ (props changed)
branches/frame_order_cleanup/auto_analyses/relax_disp_repeat_cpmg.py
Propchange: branches/frame_order_cleanup/
------------------------------------------------------------------------------
--- svnmerge-integrated (original)
+++ svnmerge-integrated Thu Oct 2 13:08:52 2014
@@ -1 +1 @@
-/trunk:1-26092
+/trunk:1-26092,26095
Modified: branches/frame_order_cleanup/auto_analyses/relax_disp_repeat_cpmg.py
URL:
http://svn.gna.org/viewcvs/relax/branches/frame_order_cleanup/auto_analyses/relax_disp_repeat_cpmg.py?rev=26131&r1=26130&r2=26131&view=diff
==============================================================================
--- branches/frame_order_cleanup/auto_analyses/relax_disp_repeat_cpmg.py
(original)
+++ branches/frame_order_cleanup/auto_analyses/relax_disp_repeat_cpmg.py
Thu Oct 2 13:08:52 2014
@@ -34,7 +34,7 @@
from datetime import datetime
from glob import glob
from os import F_OK, access, getcwd, sep
-from numpy import asarray, arange, concatenate, max, mean, min, sqrt, std,
sum
+from numpy import any, asarray, arange, concatenate, max, mean, min, sqrt,
std, sum
if dep_check.scipy_module:
from scipy.stats import pearsonr
import sys
@@ -1715,6 +1715,7 @@
for param in params_list:
# Store in list.
res_dic[str(glob_ini)]['params'][param] = []
+ res_dic[str(glob_ini)]['params'][param+'_resi'] = []
# Prepare to store individual per spin.
res_dic[str(glob_ini)][param] = {}
@@ -1746,12 +1747,16 @@
res_dic[str(glob_ini)]['params'][param].append(param_val)
res_dic[str(glob_ini)][param][spin_id][param_key] = param_val
+
res_dic[str(glob_ini)]['params'][param+'_resi'].append("%i_%3.0f"%(resi,
frq/1E6))
+
else:
# Get the value.
param_val = deepcopy(getattr(cur_spin, param))
# Store in list and per spin.
res_dic[str(glob_ini)]['params'][param].append(param_val)
res_dic[str(glob_ini)][param][spin_id][param_key] =
param_val
+
+
res_dic[str(glob_ini)]['params'][param+'_resi'].append("%i"%(resi))
# Print
subtitle(file=sys.stdout, text="The minimised valus for '%s'
model for pipe='%s at %s'" % (model, pipe_name, glob_ini), prespace=3)
@@ -1800,6 +1805,16 @@
x = data_x[str(glob_ini_x)]['params'][param]
y = data_y[str(glob_ini_y)]['params'][param]
+ # Relative uncertainty / fractional uncertainty / precision
+ precision = abs(y-x) / ((x+y)/2)
+ # Count outliers. Value differ more than than the value
itself.
+ precision_outlier = precision > 1.00
+ #precision_outlier = precision > 0.020
+ precision_outlier_nr = sum(precision_outlier)
+ resis = data_x[str(glob_ini_x)]['params'][param+'_resi']
+ resi_outlier_arr = asarray(resis)[precision_outlier]
+ resi_outlier_arr_str = ":"+','.join(str(e) for e in
resi_outlier_arr)
+
np = len(y)
# Linear a, with no intercept.
@@ -1847,7 +1862,10 @@
y_k = y[k]
x_arange_k = x_arange[k]
y_arange_k = y_arange[k]
- data_dic[method_xy_NI].append(["%3.5f"%x_k, "%3.5f"%y_k,
"%3.5f"%x_arange_k, "%3.5f"%y_arange_k])
+ precision_k = precision[k]
+ resi_k = resis[k]
+
+ data_dic[method_xy_NI].append(["%3.5f"%x_k, "%3.5f"%y_k,
"%3.5f"%x_arange_k, "%3.5f"%y_arange_k, "%3.5f"%precision_k, "%s"%resi_k,
"%i"%precision_outlier_nr, "%s"%resi_outlier_arr_str])
plt.tight_layout()
@@ -1874,7 +1892,7 @@
method_y_NI = "%s_%s_%s%s" % (analysis, param, method_y,
glob_ini_y)
method_x_NI_lin = "%s_%s_lin_%s%s" % (analysis, param,
method_x, glob_ini_x)
method_y_NI_lin = "%s_%s_lin_%s%s" % (analysis, param,
method_y, glob_ini_y)
- headings_j = headings_j + [method_x_NI, method_y_NI,
method_x_NI_lin, method_y_NI_lin]
+ headings_j = headings_j + [method_x_NI, method_y_NI,
method_x_NI_lin, method_y_NI_lin, "abs_diff_frac", "resi", "outlier_nr",
"outlier_resi"]
method_xy_NI = "%s_%s_%s%s_%s%s" % (analysis, param,
method_x, glob_ini_x, method_y, glob_ini_y)
method_xy_NI_j.append(method_xy_NI)
@@ -1901,6 +1919,7 @@
data_param_row = data_param[k]
except IndexError:
data_param_row = len(data_param[0]) * ['0.0']
+ data_param_row[-1] = ':'
data_row = data_row + data_param_row
@@ -1981,6 +2000,8 @@
# Other stats.
res_dic[method_cur][param]['r_xy'] = []
res_dic[method_cur][param]['a'] = []
+ res_dic[method_cur][param]['precision_outlier_nr'] = []
+ res_dic[method_cur][param]['resi_outlier'] = []
# Now loop over glob_ini:
for glob_ini in list_glob_ini:
@@ -1990,6 +2011,7 @@
# Get the data.
param_arr = min_dic[str(glob_ini)]['params'][param]
+ resis = min_dic[str(glob_ini)]['params'][param+'_resi']
# This require that all number of points are equal.
# If they are not of same length, then dont even bother
to continue.
@@ -2017,19 +2039,34 @@
a = sum(x*y) / sum(x**2)
r_xy = sum(x*y) / sqrt(sum(x**2) * sum(y**2))
- print(param, method_ref, method_cur,
sampling_sparseness, glob_ini, r_xy**2, a)
+ # Relative uncertainty / fractional uncertainty /
precision
+ precision = abs(y-x) / ((x+y)/2)
+ # Count outliers. Value differ more than than the value
itself.
+ precision_outlier = precision > 1.00
+ #precision_outlier = precision > 0.02
+ precision_outlier_nr = sum(precision_outlier)
+
+ resi_outlier_arr = asarray(resis)[precision_outlier]
+ resi_outlier_arr_str = ":"+','.join(str(e) for e in
resi_outlier_arr)
+
+ print(param, method_ref, method_cur,
sampling_sparseness, glob_ini, r_xy**2, a, precision_outlier_nr,
resi_outlier_arr_str)
# Store to result dic.
res_dic[method_cur][param][str(glob_ini)]['r_xy'] = r_xy
res_dic[method_cur][param]['r_xy'].append(r_xy)
res_dic[method_cur][param][str(glob_ini)]['a'] = a
res_dic[method_cur][param]['a'].append(a)
+
res_dic[method_cur][param][str(glob_ini)]['precision_outlier_nr'] =
precision_outlier_nr
+
res_dic[method_cur][param]['precision_outlier_nr'].append(precision_outlier_nr)
+
res_dic[method_cur][param]['resi_outlier'].append(resi_outlier_arr_str)
res_dic[method_cur][param]['sampling_sparseness'] =
asarray(res_dic[method_cur][param]['sampling_sparseness'])
res_dic[method_cur][param]['glob_ini'] =
asarray(res_dic[method_cur][param]['glob_ini'])
res_dic[method_cur][param]['r_xy'] =
asarray(res_dic[method_cur][param]['r_xy'])
res_dic[method_cur][param]['a'] =
asarray(res_dic[method_cur][param]['a'])
+ res_dic[method_cur][param]['precision_outlier_nr'] =
asarray(res_dic[method_cur][param]['precision_outlier_nr'])
+ res_dic[method_cur][param]['resi_outlier'] =
asarray(res_dic[method_cur][param]['resi_outlier'])
return res_dic
@@ -2074,7 +2111,7 @@
SS = min_stat_dic[method][param]['sampling_sparseness']
# Add to headings.
- headings = headings + ['method_%s'%param, 'SS', 'NI',
'slope', 'rxy2']
+ headings = headings + ['method_%s'%param, 'SS', 'NI',
'slope', 'rxy2', 'outlier_nr', 'resi_outlier']
# Get stats.
# Linear regression slope, without intercept
@@ -2093,13 +2130,19 @@
max_r_xy2 = max(r_xy2)
if min(r_xy2) < min_r_xy2:
min_r_xy2 = min(r_xy2)
+
+ # Get the precision_outlier_nr, where values change more
than the value itself.
+ precision_outlier_nr =
min_stat_dic[method][param]['precision_outlier_nr']
+ resi_outlier = min_stat_dic[method][param]['resi_outlier']
# Add to data.
for i, NI_i in enumerate(NI):
SS_i = SS[i]
a_i = a[i]
r_xy2_i = r_xy2[i]
- data_dic[method][param][str(i)] = ["%3.5f"%SS_i,
"%i"%NI_i, "%3.5f"%a_i, "%3.5f"%r_xy2_i]
+ precision_outlier_nr_i = precision_outlier_nr[i]
+ resi_outlier_i = resi_outlier[i]
+ data_dic[method][param][str(i)] = ["%3.5f"%SS_i,
"%i"%NI_i, "%3.5f"%a_i, "%3.5f"%r_xy2_i, "%i"%precision_outlier_nr_i,
"%s"%resi_outlier_i]
if i > i_max:
i_max = i
@@ -2125,7 +2168,7 @@
if str(i) in data_dic_m[param]:
data_i = data_i + ["%s_%s" % (method, param)] +
data_dic_m[param][str(i)]
else:
- data_i = data_i + ["%s_%s" % (method, param)] +
["0", "0", "0", "0"]
+ data_i = data_i + ["%s_%s" % (method, param)] +
["0", "0", "0", "0", "0", ":"]
data.append(data_i)
r26131 - in /branches/frame_order_cleanup: ./ auto_analyses/relax_disp_repeat_cpmg.py