推荐算法评估准确率，召回率

Question

在做评估模型召回率，准确率，F1时出现错误，说是括号问题，查了很多不知道哪里的问题

def evaluate_recommendations(test_set, userma_to_num, data_handling, predict, k=10):
    TP = 0  # True Positives
    FP = 0  # False Positives
    FN = 0  # False Negatives
    for user_id in test_set.index:
# Generate recommendations
        recommendations = recommend_movies(user_id, userma_to_num, data_handling, predict, k)
# Get actual liked movies from test set
        actual_liked_movies = test_set.loc[user_id][test_set.loc[user_id] > 0].index.tolist()
        for movie_id in recommendations:
            if movie_id in actual_liked_movies:
                TP += 1
            else:
                FP += 1
        for movie_id in actual_liked_movies:
            if movie_id not in recommendations:
                FN += 1
        precision = TP / (TP + FP)
        recall = TP / (TP + FN)
        f1_score = 2 * (precision * recall) / (precision + recall)
    return precision, recall, f1_score
def recommend_movies(user_id, userma_to_num, data_handling, predict, k=10):
    a = predict[:, user_id]
    max_indexs = heapq.nlargest(k, range(len(a)), a.take)
    return max_indexs

def main():
# Your existing code here...
# Splitting data into training and testing sets (80% train, 20% test)
    train_set = df1.sample(frac=0.8, random_state=42)
    test_set = df1.drop(train_set.index)
# Preprocessing...
# Training the model...
# Generate recommendations and evaluate
    
    precision, recall, f1_score = evaluate_recommendations(test_set, userma_to_num, data_handling, predict)

    print("Precision:", precision)
    print("Recall:", recall)
    print("F1 Score:", f1_score)
if __name__ == "__main__":
    main()

Answer 1

m0_56062032 中午好🌞🌞🌞
本答案参考ChatGPT-3.5

你遇到的错误是括号问题，可能是因为计算公式的括号使用不正确导致。根据你提供的代码和问题描述，括号问题可能出现在计算precision、recall和f1_score的公式中。

为了解决这个问题，你可以通过调整括号的位置来确保计算顺序正确。下面是一些解决方案：

1. 检查公式计算的括号位置：检查precision、recall和f1_score的公式中括号的位置是否正确。确保括号位于正确的位置以确保计算的正确性。

2. 计算之前先重置计数器：在每次循环开始之前，重置TP、FP和FN的计数器为0。这样可以确保每次循环的计数结果是正确的，而不会受到上一次循环的影响。

3. 检查recommend_movies函数的输出格式：确保recommend_movies函数返回的结果是一个包含影片编号的列表。如果格式不正确，可能会导致recommendations列表中的影片不能正确地和actual_liked_movies列表中的影片进行匹配。

4. 检查数据处理和预测函数的正确性：确保data_handling和predict函数返回正确的结果。这些函数可能会影响到recommend_movies函数的结果，进而影响到计算推荐评估指标的正确性。

修改后的代码如下所示：

def evaluate_recommendations(test_set, userma_to_num, data_handling, predict, k=10):
    TP = 0  # True Positives
    FP = 0  # False Positives
    FN = 0  # False Negatives
    for user_id in test_set.index:
        # Generate recommendations
        recommendations = recommend_movies(user_id, userma_to_num, data_handling, predict, k)
        # Get actual liked movies from test set
        actual_liked_movies = test_set.loc[user_id][test_set.loc[user_id] > 0].index.tolist()
        for movie_id in recommendations:
            if movie_id in actual_liked_movies:
                TP += 1
            else:
                FP += 1
        for movie_id in actual_liked_movies:
            if movie_id not in recommendations:
                FN += 1
        precision = TP / (TP + FP)
        recall = TP / (TP + FN)
        f1_score = 2 * (precision * recall) / (precision + recall)
        
    return precision, recall, f1_score

def recommend_movies(user_id, userma_to_num, data_handling, predict, k=10):
    a = predict[:, user_id]
    max_indexs = heapq.nlargest(k, range(len(a)), a.take)
    return max_indexs

def main():
    # Your existing code here...
    # Splitting data into training and testing sets (80% train, 20% test)
    train_set = df1.sample(frac=0.8, random_state=42)
    test_set = df1.drop(train_set.index)
    # Preprocessing...
    # Training the model...
    # Generate recommendations and evaluate
    precision, recall, f1_score = evaluate_recommendations(test_set, userma_to_num, data_handling, predict)

    print("Precision:", precision)
    print("Recall:", recall)
    print("F1 Score:", f1_score)

if __name__ == "__main__":
    main()

请确保在运行代码之前，对test_set、userma_to_num、data_handling和predict变量进行正确的赋值和初始化。