具体的代码实现是怎么样的呢，基于AVEC2014数据集

基于AVEC2014数据集的主动学习抑郁症辅助诊断系统的代码实现比较复杂，需要进行数据预处理、特征提取、模型训练等多个步骤。

下面是一个简单的代码示例，用于说明如何使用Python和Scikit-learn库来实现基于主动学习的抑郁症辅助诊断系统：

import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report
from sklearn.utils import shuffle

# Load data
data = pd.read_csv('avec2014.csv')

# Preprocess data
data = data.dropna(subset=['transcription'])
data = shuffle(data)

# Split data into train and test sets
train_data, test_data = train_test_split(data, test_size=0.2)

# Define active learning function
def active_learning(X_pool, y_pool, X_test, y_test, n_rounds=10, n_query=10, model=None):
    # Initialize model
    if model is None:
        model = LogisticRegression()

    # Initialize active learning loop
    for i in range(n_rounds):
        # Train model on labeled data
        model.fit(X_pool, y_pool)

        # Predict on test data
        y_pred = model.predict(X_test)

        # Evaluate model performance
        report = classification_report(y_test, y_pred, output_dict=True)
        acc = report['accuracy']
        print('Round %d Accuracy: %.4f' % (i+1, acc))

        # Query for new samples
        idx = np.argpartition(model.predict_proba(X_pool)[:, 1], -n_query)[-n_query:]
        X_new, y_new = X_pool[idx], y_pool[idx]

        # Remove queried samples from pool
        X_pool, y_pool = np.delete(X_pool, idx, axis=0), np.delete(y_pool, idx)

        # Add queried samples to labeled data
        X_labeled, y_labeled = np.concatenate([X_new, X_pool]), np.concatenate([y_new, y_pool])

    return model

# Extract features
vectorizer = TfidfVectorizer(stop_words='english', max_features=1000)
X = vectorizer.fit_transform(train_data['transcription'])
y = train_data['label']

# Split data into labeled and unlabeled sets
X_labeled, X_unlabeled, y_labeled, y_unlabeled = train_test_split(X, y, test_size=0.9)

# Train model using active learning
model = active_learning(X_unlabeled, y_unlabeled, X_test=X, y_test=y, n_rounds=10, n_query=10)

# Evaluate model performance on test data
X_test = vectorizer.transform(test_data['transcription'])
y_test = test_data['label']
y_pred = model.predict(X_test)
report = classification_report(y_test, y_pred)
print(report)

这个代码示例使用了一个简单的文本分类模型（逻辑回归），使用TF-IDF算法进行特征提取，并使用主动学习算法选择最具代表性和区分度的样本进行训练。需要注意的是，这个代码示例仅用于说明如何使用Scikit-learn库实现主动学习算法，实际应用中需要根据具体的数据和场景进行修改和扩展。