```python # model description model_spec = "svm_y-effort_x-choiceS" ycol = "effort" Xcols = ["choiceS"] def fit_classifier(X, y, loss="hinge", penalty="l2", cv=5, alpha=0.3): model = SGDClassifier(loss=loss, penalty=penalty, alpha=alpha) return cross_val_score(model, X, y, cv=cv).mean() ``` Low overall decoding accuracy (about 60-63%) ```r > dt_scores[include == 1, mean(score, na.rm = T), by = group] group V1 1: charity 0.6353201 2: otherperson 0.6260444 3: self 0.6052593 ``` Weird correlations ![[Pasted image 20201004221841.png]]