我目前正在使用名为LightFM的Python库.但是我在将交互传递给fit()方法时遇到了一些麻烦.
Python版本:3库:http://lyst.github.io/lightfm/docs/lightfm.html
文档说明我应该创建一个以下类型的稀疏矩阵:interaction(np.float32 coo_matrix of shape [n_users,n_items]) - 矩阵
但我似乎无法使它工作,它始终建议相同...
更新:当执行它时,top_items变量说出以下内容,无论它迭代哪个用户而不是任何其他项目(牛肉或沙拉),所以看起来我做错了.它每次输出:['Cake''Cheese']
这是我的代码:
import numpy as np
from lightfm.datasets import fetch_movielens
from lightfm import LightFM
from scipy.sparse import coo_matrix
import scipy.sparse as sparse
import scipy
// Users, items
data = [
[1, 0],
[2, 1],
[3, 2],
[4, 3]
]
items = np.array(["Cake", "Cheese", "Beef", "Salad"])
data = coo_matrix(data)
#create model
model = LightFM(loss='warp')
#train model
model.fit(data, epochs=30, num_threads=2)
// Print training data
print(data)
def sample_recommendation(model, …Run Code Online (Sandbox Code Playgroud) python recommendation-engine machine-learning scipy sparse-matrix