# AFM
X = hstack((S, Q, O))
y = np.array(y)
l2 = [1.0 for i in range(S.shape[1])]
l2 += [0.0 for i in range(Q.shape[1])]
l2 += [0.0 for i in range(O.shape[1])]
bounds = [(None, None) for i in range(S.shape[1])]
bounds += [(None, None) for i in range(Q.shape[1])]
bounds += [(0, None) for i in range(O.shape[1])]
X = X.toarray()
X2 = Q.toarray()
if args.model == "AFM":
m = CustomLogistic(bounds=bounds, l2=l2, fit_intercept=False)
m.fit(X, y)
yHat = 1 - m.predict_proba(X)
elif args.model == "AFM+S":
m = BoundedLogistic(first_bounds=bounds, first_l2=l2)
m.fit(X, X2, y)
yHat = 1 - m.predict_proba(X, X2)
else:
raise ValueError("Model type not supported")
headers = original_headers + ["Predicted Error Rate (%s)" % kc_model]
outfilePath = args.workingDir + "/output.txt"
outfile = open(outfilePath, 'w')
outfile.write("\t".join(headers) + "\n")
cntRowMissOpp = 0
for i, row in enumerate(original_step_data):
# AFM
X = hstack((S, Q, O))
y = np.array(y)
l2 = [1.0 for i in range(S.shape[1])]
l2 += [0.0 for i in range(Q.shape[1])]
l2 += [0.0 for i in range(O.shape[1])]
bounds = [(None, None) for i in range(S.shape[1])]
bounds += [(None, None) for i in range(Q.shape[1])]
bounds += [(0, None) for i in range(O.shape[1])]
X = X.toarray()
X2 = Q.toarray()
if args.m == "AFM":
m = CustomLogistic(bounds=bounds, l2=l2, fit_intercept=False)
m.fit(X, y)
coef_s = m.coef_[0:S.shape[1]]
coef_s = [[k, v, invlogit(v)]
for k, v in sv.inverse_transform([coef_s])[0].items()]
coef_q = m.coef_[S.shape[1]:S.shape[1] + Q.shape[1]]
coef_qint = qv.inverse_transform([coef_q])[0]
coef_o = m.coef_[S.shape[1] + Q.shape[1]:S.shape[1] + Q.shape[1] +
O.shape[1]]
coef_qslope = ov.inverse_transform([coef_o])[0]
kc_vals = []
all_kcs = set(coef_qint).union(set(coef_qslope))
for kc in all_kcs:
kc_vals.append([
kc,
def afm(kcs,
opps,
actuals,
stu,
student_label,
item_label,
nfolds=3,
seed=None):
"""
Executes AFM on the provided data and returns model fits and parameter estimates
"""
sv = DictVectorizer()
qv = DictVectorizer()
ov = DictVectorizer()
S = sv.fit_transform(stu)
Q = qv.fit_transform(kcs)
O = ov.fit_transform(opps)
X = hstack((S, Q, O))
y = np.array(actuals)
l2 = [1.0 for i in range(S.shape[1])]
l2 += [0.0 for i in range(Q.shape[1])]
l2 += [0.0 for i in range(O.shape[1])]
bounds = [(None, None) for i in range(S.shape[1])]
bounds += [(None, None) for i in range(Q.shape[1])]
bounds += [(0, None) for i in range(O.shape[1])]
X = X.toarray()
X2 = Q.toarray()
model = CustomLogistic(bounds=bounds, l2=l2, fit_intercept=False)
model.fit(X, y)
coef_s = model.coef_[0:S.shape[1]]
coef_s = [[k, v, invlogit(v)]
for k, v in sv.inverse_transform([coef_s])[0].items()]
coef_q = model.coef_[S.shape[1]:S.shape[1] + Q.shape[1]]
coef_qint = qv.inverse_transform([coef_q])[0]
coef_o = model.coef_[S.shape[1] + Q.shape[1]:S.shape[1] + Q.shape[1] +
O.shape[1]]
coef_qslope = ov.inverse_transform([coef_o])[0]
kc_vals = []
all_kcs = set(coef_qint).union(set(coef_qslope))
for kc in all_kcs:
kc_vals.append([
kc,
coef_qint.setdefault(kc, 0.0),
invlogit(coef_qint.setdefault(kc, 0.0)),
coef_qslope.setdefault(kc, 0.0)
])
cvs = [
KFold(len(y), n_folds=nfolds, shuffle=True, random_state=seed),
StratifiedKFold(y, n_folds=nfolds, shuffle=True, random_state=seed),
LabelKFold(student_label, n_folds=nfolds),
LabelKFold(item_label, n_folds=nfolds)
]
scores = []
for cv in cvs:
score = []
for train_index, test_index in cv:
X_train, X_test = X[train_index], X[test_index]
y_train, y_test = y[train_index], y[test_index]
model.fit(X_train, y_train)
score.append(model.mean_squared_error(X_test, y_test))
scores.append(np.mean(np.sqrt(score)))
return scores, kc_vals, coef_s
y = np.array(y)
# Regularize the student intercepts
l2 = [1.0 for i in range(S.shape[1])]
l2 += [0.0 for i in range(Q.shape[1])]
l2 += [0.0 for i in range(O.shape[1])]
# Bound the learning rates to be positive
bounds = [(None, None) for i in range(S.shape[1])]
bounds += [(None, None) for i in range(Q.shape[1])]
bounds += [(0, None) for i in range(O.shape[1])]
X = X.toarray()
X2 = Q.toarray()
afm = CustomLogistic(bounds=bounds, l2=l2, fit_intercept=False)
afm.fit(X, y)
yAFM = afm.predict_proba(X)
afms = BoundedLogistic(first_bounds=bounds, first_l2=l2)
afms.fit(X, X2, y)
yAFMS = afms.predict_proba(X, X2)
#plotkcs = ['All Knowledge Components']
plotkcs = list(set([kc for row in kcs for kc in row])) + ['All Knowledge Components']
#f, subplots = plt.subplots(len(plotkcs))
for plot_id, plotkc in enumerate(plotkcs):
plt.figure(plot_id+1)
y = np.array(y)
# Regularize the student intercepts
l2 = [1.0 for i in range(S.shape[1])]
l2 += [0.0 for i in range(Q.shape[1])]
l2 += [0.0 for i in range(O.shape[1])]
# Bound the learning rates to be positive
bounds = [(None, None) for i in range(S.shape[1])]
bounds += [(None, None) for i in range(Q.shape[1])]
bounds += [(0, None) for i in range(O.shape[1])]
X = X.toarray()
X2 = Q.toarray()
afm = CustomLogistic(bounds=bounds, l2=l2, fit_intercept=False)
afm.fit(X, y)
yAFM = afm.predict_proba(X)
afms = BoundedLogistic(first_bounds=bounds, first_l2=l2)
afms.fit(X, X2, y)
yAFMS = afms.predict_proba(X, X2)
#plotkcs = ['All Knowledge Components']
plotkcs = list(set([kc for row in kcs for kc in row])) + ['All Knowledge Components']
#f, subplots = plt.subplots(len(plotkcs))
for plot_id, plotkc in enumerate(plotkcs):
plt.figure(plot_id+1)