gistfile1.txt

depth = 10
seed_range = range(0, 3000,288)
acc_vs_seed_result = {"seed": [],\
        "train_acc": [],
        "valid_acc": [],
        "top_feature": [],
        "second_feature": [],
        "third_feature": []}

for seed in seed_range:
    model = H2ORandomForestEstimator(model_id="model", \
                                       sample_rate=0.7, \
                                       ntrees=1, \
                                       max_depth=depth, \
                                       seed=seed)
    model.train(x=x, y=y, training_frame=train)
    predict_valid = model.predict(valid[x])
    predict_train = model.predict(train[x])
    acc_vs_seed_result["seed"].append(seed)
    acc_vs_seed_result["valid_acc"].append((predict_valid["predict"] == valid["Survived"]).mean()[0])
    acc_vs_seed_result["train_acc"].append((predict_train["predict"] == train["Survived"]).mean()[0])
    acc_vs_seed_result["top_feature"].append(model.varimp()[0][0])
    acc_vs_seed_result["second_feature"].append(model.varimp()[1][0])
    acc_vs_seed_result["third_feature"].append(model.varimp()[2][0])
    
acc_vs_seed_result_df = pd.DataFrame(acc_vs_seed_result)
cols = ["seed", "train_acc", "valid_acc", "top_feature", "second_feature", "third_feature"]
acc_vs_seed_result_df = acc_vs_seed_result_df[cols]
acc_vs_seed_result_df