flattened.extend(s)
flattened = DataFrame({
"text": flattened,
"group": list(fr.group) * len(columns)
})
if local:
l = [
leave_one_out(fr.spatial, fr.group, "spatial_loo_svm"),
leave_one_out(fr.social, fr.group, "social_loo_svm"),
leave_one_out(fr.hypothetical, fr.group, "hypothetical_loo_svm"),
leave_one_out(fr.control, fr.group, "control_loo_svm"),
leave_one_out(fr.temporal, fr.group, "temporal_loo_svm"),
leave_one_out(flattened.text, flattened.group, "all_loo_svm")
]
print(l)
df = DataFrame(l)
df.to_csv("classifier_performance.csv")
else:
import sbatch
sbatch.load("leave_one_out",
[fr.temporal, fr.group, "temporal_loo_svm"])
sbatch.load("leave_one_out", [fr.spatial, fr.group, "spatial_loo_svm"])
sbatch.load("leave_one_out", [fr.social, fr.group, "social_loo_svm"])
sbatch.load("leave_one_out",
[fr.hypothetical, fr.group, "hypothetical_loo_svm"])
sbatch.load("leave_one_out", [fr.control, fr.group, "control_loo_svm"])
sbatch.load("leave_one_out",
[flattened.text, flattened.group, "all_loo_svm"])
sbatch.launch()
fr.control,
fr.temporal
]
names = [
"spatial",
"social",
"hypothetical",
"control",
"temporal"
]
get_shared_features(dims, names, fr.group).to_csv("features_for_each_domain.csv")
get_shared_features(dims, names, fr.group, across=True).to_csv("shared_features.csv")
if local:
outs = []
for i1 in range(len(dims)):
for i2 in range(len(dims)):
outs.append(cross_decode(dims[i1], fr.group, dims[i2], fr.group, names[i1] + "_to_" + names[i2]))
df = DataFrame(outs)
df.to_csv("cross_encoding_performance.csv")
else:
import sbatch
for i1 in range(len(dims)):
for i2 in range(len(dims)):
sbatch.load("cross_decode", [dims[i1], fr.group, dims[i2], fr.group, names[i1] + "_to_" + names[i2]])
sbatch.launch()
print("jobs submitted to the cluster. Run the collation scripts when they're done!")