瑶子的博客

import numpy as npimport matplotlib.pyplot as pltimport math#老虎机个数number_of_bandits=10#老虎机的臂数number_of_arms=10#尝试数number_of_pulls=10000#epsepsilon=0.3#最小的decaymin_temp = 0.1#衰减率decay_rat...

1、导入库文件from __future__ import print_functionimport collectionsimport mathimport numpy as npimport osimport randomimport tensorflow as tfimport zipfilefrom matplotlib import pylabfrom six.mov...

# Copyright 2015 The TensorFlow Authors. All Rights Reserved.## Licensed under the Apache License, Version 2.0 (the "License");# you may not use this file except in compliance with the License.# Y...

1 learning to rank 回顾Learning to Rank for Information Retrieval，Tie-Yan ：Liuhttps://www.cda.cn/uploadfile/image/20151220/20151220115436_46293.pdf2、点击预估模型模型：

业界通用推荐系统框架结构如下：Match & Rank定义：Match基于当前user(profile、history)和context,快速在全库里找到TopN最相关的item，给Rank来做小范围综合多目标最大化通常做法：用各种算法做召回，比如user/item/model-based CF,Content-based,Demographic-based,DNN-Embeddi...

一、多目标排序1.1为什么要有多目标排序？⼯业界推荐系统多基于隐式反馈•Global bias：不同⽬标表达不同的偏好程度•Item bias：单个⽬标衡量不全⾯（标题党）•User bias：⽤户表达满意度的⽅式不同•综合⽬标收益最⼤化多⽬标排序的难点多⽬标 vs CTR预估**•**部分⽬标数据稀疏，模型准确率低**•**在线服务计算量⼤**•**多个⽬标间重要性难以量...