def print_dict_latex(stats, top, patient):
print("\\begin{table}[ht]")
print("\\resizebox{\\textwidth}{!}{\\begin{tabular}{l|| *{6}{c|}}")
print(
"\\hline \\hline",
"\n\\textbf{Models} & \\textbf{RMSE} & \\textbf{MAPE} & \\textbf{TG} & \\textbf{AP} "
"& \\textbf{BE} & \\textbf{EP} \\\\ \n\\hline \\hline")
cpt = 0
for key in stats.keys():
mean = [stats[key][key_1][0] for key_1 in stats[key].keys()]
diff = [stats[key][key_1][1] for key_1 in stats[key].keys()]
if cpt == 0:
print_latex(mean, diff, key)
print("\\hline")
else:
print_latex(mean, diff, str(cpt) + "/" + " glucose + " + key)
cpt += 1
print("\\hline \\hline")
print("\\end{tabular}}")
if patient is None:
print("\\caption{\\detokenize{Top 5", top, "global models}}")
else:
print("\\caption{\\detokenize{Top 5", top, "models for patient",
str(patient) + "}}")
print("\\end{table}")
def to_latex(self, table="acc", model_name=None):
"""
Format the results into a string for the paper in LATEX
:param table: either "acc" or "cg_ega", corresponds to the table
:param model_name: prefix of the string, name of the model
:return:
"""
mean, std = self.compute_results()
if table == "cg_ega":
keys = [
"CG_EGA_AP_hypo", "CG_EGA_BE_hypo", "CG_EGA_EP_hypo",
"CG_EGA_AP_eu", "CG_EGA_BE_eu", "CG_EGA_EP_eu",
"CG_EGA_AP_hyper", "CG_EGA_BE_hyper", "CG_EGA_EP_hyper"
]
mean = [mean[k] * 100 for k in keys]
std = [std[k] * 100 for k in keys]
elif table == "general":
acc_keys = ["RMSE", "MAPE", "CG_EGA_AP", "CG_EGA_BE", "CG_EGA_EP"]
mean = [
mean[k] if k not in ["CG_EGA_AP", "CG_EGA_BE", "CG_EGA_EP"]
else mean[k] * 100 for k in acc_keys
]
std = [
std[k] if k not in ["CG_EGA_AP", "CG_EGA_BE", "CG_EGA_EP"] else
std[k] * 100 for k in acc_keys
]
print_latex(mean, std, label=self.model)
def erroneous_prediction_analysis(self, dataset, disp=False):
"""
Compute the responsability of P-EGA or R-EGA on EP predictions for all the subjects of given dataset
:param dataset: name of dataset (e.g., "ohio")
:param disp: if True, print latex format of DataFrame
:return:
"""
cg_ega_analysis = []
for i, sbj, in enumerate(misc.datasets.datasets[dataset]["subjects"]):
mean_sbj, std_sbj = self.erroneous_prediction_analysis_per_subject(
dataset, sbj)
cg_ega_analysis.append(mean_sbj)
if disp:
print_latex(mean_sbj, std_sbj, label=sbj)
cg_ega_analysis = pd.DataFrame(
data=cg_ega_analysis,
columns=[
"Hypo_EP_P", "Hypo_EP_R", "Hypo_EP_B", "Eu_EP_P", "Eu_EP_R",
"Eu_EP_B", "Hyper_EP_P", "Hyper_EP_R", "Hyper_EP_B"
],
index=misc.datasets.datasets[dataset]["subjects"])
mean = cg_ega_analysis.mean(axis=0)
std = cg_ega_analysis.std(axis=0)
cg_ega_analysis.loc["mean"] = mean
cg_ega_analysis.loc["std"] = std
if disp:
print_latex(mean, std, label=dataset)
return cg_ega_analysis
def to_latex(self, iter=0, table="general", model_name=None):
"""
Format the results into a string for the paper in LATEX
:param table: either "acc" or "cg_ega", corresponds to the table
:param model_name: prefix of the string, name of the model
:return:
"""
mean, std = self.compute_results_iter(iter)
if table == "p_ega":
p_ega_keys = [
"P_EGA_A+B", "P_EGA_A", "P_EGA_B", "P_EGA_C", "P_EGA_D",
"P_EGA_E"
]
mean = [mean[k] * 100 for k in p_ega_keys]
std = [std[k] * 100 for k in p_ega_keys]
elif table == "r_ega":
r_ega_keys = [
"R_EGA_A+B", "R_EGA_A", "R_EGA_B", "R_EGA_lC", "R_EGA_uC",
"R_EGA_lD", "R_EGA_uD", "R_EGA_lE", "R_EGA_uC"
]
mean = [mean[k] * 100 for k in r_ega_keys]
std = [std[k] * 100 for k in r_ega_keys]
elif table == "cg_ega":
cg_ega_keys = [
"CG_EGA_AP_hypo", "CG_EGA_BE_hypo", "CG_EGA_EP_hypo",
"CG_EGA_AP_eu", "CG_EGA_BE_eu", "CG_EGA_EP_eu",
"CG_EGA_AP_hyper", "CG_EGA_BE_hyper", "CG_EGA_EP_hyper"
]
mean = [mean[k] * 100 for k in cg_ega_keys]
std = [std[k] * 100 for k in cg_ega_keys]
elif table == "acc":
acc_keys = ["RMSE", "MAPE", "MASE", "TG"]
mean = [mean[k] for k in acc_keys]
std = [std[k] for k in acc_keys]
elif table == "general":
acc_keys = [
"RMSE", "MAPE", "MASE", "TG", "CG_EGA_AP", "CG_EGA_BE",
"CG_EGA_EP"
]
mean = [
mean[k] if k not in ["CG_EGA_AP", "CG_EGA_BE", "CG_EGA_EP"]
else mean[k] * 100 for k in acc_keys
]
std = [
std[k] if k not in ["CG_EGA_AP", "CG_EGA_BE", "CG_EGA_EP"] else
std[k] * 100 for k in acc_keys
]
print_latex(mean, std, label=self.model)
def print_dict_stats_physio(stats, variable, patient=None):
print("\\begin{table}[ht]")
print("\\resizebox{\\textwidth}{!}{\\begin{tabular}{l|| *{6}{c|}}")
print(
"\\hline \\hline",
"\n\\textbf{Features impact} & \\textbf{RMSE} & \\textbf{MAPE} & \\textbf{TG} & "
"\\textbf{AP} & \\textbf{BE} & \\textbf{EP} \\\\ \n\\hline")
for key in stats.keys():
mean = [stats[key][key_1] for key_1 in stats[key].keys()]
diff = [stats[key][key_1] for key_1 in stats[key].keys()]
print_latex(mean, diff, " vs ".join(key))
print("\\hline \\hline")
print("\\end{tabular}}")
if patient is None:
print("\\caption{" + variable, "impact on all the patients}")
else:
print("\\caption{" + variable, "impact for patient",
str(patient) + "}")
print("\\end{table}")
