基于机器学习的关键词竞价系统的研究

发布时间：2018-03-31 09:33

本文选题：搜索引擎营销　切入点：关键词竞价　出处：《辽宁科技大学》2013年硕士论文

【摘要】：随着近年来互联网的迅速发展，网络广告也得到了较快的发展。当前，搜索竞价广告是当前互联网提供的主要的网络广告投放方式和最有效的营销手段，可以帮助企业借助于搜索引擎的功能和作用，寻找潜在客户，推广和销售企业的产品，为企业带来利润，许多企业都将关键词广告视为一种重要的电子商务营销方式。搜索竞价广告关键词优化是当今搜索竞价广告领域的研究热点和难点，它包含了两个方面，一方面包括优化选择关键词，，另一方面包括对所购买和投放的关键词优化出价。目前业界对购买的搜索竞价词竞价，主要是基于手动，也就是人工凭借经验来调整购买的投放关键词的出价，这无疑需要大量的时间和精力。本文的主要工作如下： 1.构建了一个商品导航的网站的同时，构建了一个聚焦爬虫系统，采集当前互联网著名的电子商务网站的商品信息，抽取后存储到数据库。 2.应用逻辑回归模型预测关键词的转化率。 3.提出三种控制方式，即：每天效果数、每天消耗额，每效果消耗额，通过实验证明了控制方式的有效性。 4.利用模型预测的关键词转化率，控制方式等，提出关键词价格计算公式，通过在百度购买关键词并推广购物导航网站，利用关键词出价系统计算关键词的价格并通过百度API进行竞价。
[Abstract]:With the rapid development of the Internet in recent years, the online advertising has also been developed rapidly. At present, the search auction advertising is the main way of Internet advertising and the most effective means of marketing. It can help enterprises to find potential customers, promote and sell their products with the help of the functions and functions of search engines, and bring profit to enterprises. Many enterprises regard keyword advertising as an important electronic commerce marketing method. Keyword optimization in search bidding advertising is a hot and difficult point in the field of search auction advertising, which contains two aspects. On the one hand, it includes optimizing the selection keywords, on the other hand, it includes optimizing the bid for the keywords purchased and put in. That is, manual use of experience to adjust the purchase of the keyword bid, which undoubtedly requires a lot of time and effort. The main work of this article is as follows:. 1. At the same time, a focused crawler system is constructed, which collects the commodity information of the famous e-commerce websites on the Internet and stores it to the database. 2. The logical regression model is used to predict the conversion rate of keywords. 3. Three kinds of control methods are put forward, that is, the number of effects per day, the amount of daily consumption and the amount of consumption per day. The effectiveness of the control method is proved by experiments. 4.Use the keyword conversion rate and control mode predicted by the model, put forward the key word price calculation formula, by buying keywords in Baidu and popularizing shopping navigation website, Use keyword bid system to calculate the price of key words and bid through Baidu API.
【学位授予单位】：辽宁科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【参考文献】