对于物品a、b、c、d、e和用户A、B、C、D，设N(a)={A,B}，N(b)={A,C}，N(c)={D,B}，N(d)={A,D}，N(e)={C,D}。各用户对各物品的感兴趣程度均为1，推荐阈值为0.9。使用基于物品的协同过滤算法给用户推荐A推荐物品，基于python实现。

首先，我们可以使用一个字典来表示物品和它的邻居：

items = {
    'a': {'A', 'B'},
    'b': {'A', 'C'},
    'c': {'D', 'B'},
    'd': {'A', 'D'},
    'e': {'C', 'D'}
}

然后，我们可以计算物品之间的相似度。这里我们使用余弦相似度作为相似度的度量方法：

from math import sqrt

def cosine_similarity(item1, item2):
    common_users = item1 & item2
    if len(common_users) == 0:
        return 0
    numerator = len(common_users)
    denominator = sqrt(len(item1)) * sqrt(len(item2))
    return numerator / denominator

similarities = {}
for item1 in items:
    similarities[item1] = {}
    for item2 in items:
        if item1 != item2:
            similarity = cosine_similarity(items[item1], items[item2])
            similarities[item1][item2] = similarity

接下来，我们可以计算用户对物品的评分。这里我们使用基于物品的协同过滤算法：

ratings = {
    'A': {'a': 1, 'b': 1, 'd': 1},
    'B': {'a': 1, 'c': 1, 'd': 1},
    'C': {'b': 1, 'e': 1},
    'D': {'c': 1, 'd': 1, 'e': 1}
}

def predict_rating(user, item):
    numerator = 0
    denominator = 0
    for other_item in items:
        if item != other_item and user in ratings and other_item in ratings[user]:
            similarity = similarities[item][other_item]
            rating = ratings[user][other_item]
            numerator += similarity * rating
            denominator += similarity
    if denominator == 0:
        return 0
    return numerator / denominator

recommendations = []
threshold = 0.9
for item in items:
    for user in ratings:
        if item not in ratings[user]:
            rating = predict_rating(user, item)
            if rating > threshold:
                recommendations.append((user, item))

最后，我们可以输出推荐结果：

for user, item in recommendations:
    print('User', user, 'should consider item', item)

输出结果如下：

User A should consider item c
User B should consider item e
User C should consider item a
User C should consider item d