展示廣告點(diǎn)擊率預(yù)估平臺(tái)的設(shè)計(jì)與實(shí)現(xiàn)
發(fā)布時(shí)間：2018-09-11 14:37
【摘要】：隨著互聯(lián)網(wǎng)產(chǎn)業(yè)的成熟以及用戶規(guī)模的擴(kuò)張,互聯(lián)網(wǎng)廣告的營銷價(jià)值也隨之不斷攀升。若能利用互聯(lián)網(wǎng)廣告的天然優(yōu)勢(shì),通過點(diǎn)擊率預(yù)估技術(shù)正確追蹤用戶對(duì)某廣告的偏好,可帶來廣告主轉(zhuǎn)化提升、用戶體驗(yàn)提升、發(fā)布者收入提升等多維度的收益。因此,本文選擇對(duì)計(jì)算廣告學(xué)中的展示廣告點(diǎn)擊率預(yù)估問題展開研究,并將在研究過程中搭建展示廣告點(diǎn)擊率預(yù)估平臺(tái),為當(dāng)前問題提供完整的機(jī)器學(xué)習(xí)解決方案。本文將系統(tǒng)地介紹展示廣告點(diǎn)擊率預(yù)估平臺(tái)的構(gòu)建過程。首先通過對(duì)點(diǎn)擊率預(yù)估問題研究意義及研究現(xiàn)狀的分析,引出本文的研究內(nèi)容。隨后,結(jié)合點(diǎn)擊率預(yù)估問題的機(jī)器學(xué)習(xí)方案步驟對(duì)點(diǎn)擊率預(yù)估問題展開業(yè)務(wù)分析,明確點(diǎn)擊率預(yù)估平臺(tái)的功能與性能需求,即針對(duì)計(jì)算廣告中的海量數(shù)據(jù),支持多種模型的獨(dú)立、混合使用,配備離線批量學(xué)習(xí)和在線學(xué)習(xí)兩種訓(xùn)練模式,為用戶提供了從特征工程、模型訓(xùn)練、模型評(píng)估、模型預(yù)測到結(jié)果分析的一站式服務(wù)。緊接著,圍繞需求分析的結(jié)果開始闡述點(diǎn)擊率預(yù)估平臺(tái)的詳細(xì)設(shè)計(jì)及實(shí)現(xiàn)過程,其中特征工程與模型訓(xùn)練的設(shè)計(jì)與實(shí)現(xiàn)是文本的研究重點(diǎn)。由于點(diǎn)擊率預(yù)估問題的數(shù)據(jù)來源往往是線上真實(shí)的服務(wù)日志,本文將通過系統(tǒng)的特征工程挖掘湮沒在大量噪聲中的有效特征,并力求使用最少的特征帶來最佳的模型預(yù)測效果。模型訓(xùn)練階段,本文選擇了適合離散高維特征場景的邏輯回歸模型,以及適合稀疏特征場景的因子分解機(jī)模型,并將其分別與上游的GBDT模型通過Stacking集成算法進(jìn)行融合,達(dá)到提升模型預(yù)估效果的目的。平臺(tái)初步實(shí)現(xiàn)后,將通過功能測試與性能測試,發(fā)現(xiàn)平臺(tái)存在的問題。通過進(jìn)一步的優(yōu)化與迭代,完成平臺(tái)的全部搭建工作。最后,對(duì)論文內(nèi)容進(jìn)行了總結(jié),并對(duì)平臺(tái)的改進(jìn)方向進(jìn)行了展望。
[Abstract]:With the maturity of the Internet industry and the expansion of the scale of users, the marketing value of Internet advertising is rising. If we can make use of the natural advantages of Internet advertising and correctly track the preference of users to a certain advertisement through the technology of predicting the click rate, we can bring about the multi-dimensional income, such as the transformation of advertisers, the improvement of user experience, the increase of publishers' income, and so on. Therefore, this paper chooses to study the prediction of display advertising click rate in computational advertising, and will build the display advertising click rate prediction platform in the process of research to provide a complete machine learning solution for current problems. This paper will systematically introduce the construction process of display advertising click rate prediction platform. Firstly, by analyzing the significance and current situation of the research on the prediction of click rate, the research content of this paper is introduced. Then, combined with the machine learning program of the click rate prediction problem, the operation analysis of the click rate prediction problem is carried out, and the function and performance requirements of the click rate prediction platform are clarified, that is, the mass data in the calculation advertising, It supports the independent, mixed use of multiple models and provides a one-stop service for users from feature engineering, model training, model evaluation, model prediction to result analysis, with offline batch learning and online learning. Then, the detailed design and implementation process of the platform is described around the results of requirement analysis, in which the design and implementation of feature engineering and model training are the focus of the text research. Because the data source of the prediction problem of click-through rate is often the online real service log, this paper will mine the valid features of annihilation in a large amount of noise through the feature engineering of the system, and make every effort to use the least features to bring the best prediction effect of the model. In the training stage of the model, the logical regression model suitable for discrete high-dimensional feature scenes and the factoring machine model for sparse feature scenes are selected, and the model is fused with the upstream GBDT model by Stacking integration algorithm, respectively. To achieve the purpose of improving the prediction effect of the model. After the initial implementation of the platform, through functional testing and performance testing, found the platform problems. Through further optimization and iteration, the platform is built. Finally, the content of the paper is summarized, and the improvement direction of the platform is prospected.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP311.52
[Abstract]:With the maturity of the Internet industry and the expansion of the scale of users, the marketing value of Internet advertising is rising. If we can make use of the natural advantages of Internet advertising and correctly track the preference of users to a certain advertisement through the technology of predicting the click rate, we can bring about the multi-dimensional income, such as the transformation of advertisers, the improvement of user experience, the increase of publishers' income, and so on. Therefore, this paper chooses to study the prediction of display advertising click rate in computational advertising, and will build the display advertising click rate prediction platform in the process of research to provide a complete machine learning solution for current problems. This paper will systematically introduce the construction process of display advertising click rate prediction platform. Firstly, by analyzing the significance and current situation of the research on the prediction of click rate, the research content of this paper is introduced. Then, combined with the machine learning program of the click rate prediction problem, the operation analysis of the click rate prediction problem is carried out, and the function and performance requirements of the click rate prediction platform are clarified, that is, the mass data in the calculation advertising, It supports the independent, mixed use of multiple models and provides a one-stop service for users from feature engineering, model training, model evaluation, model prediction to result analysis, with offline batch learning and online learning. Then, the detailed design and implementation process of the platform is described around the results of requirement analysis, in which the design and implementation of feature engineering and model training are the focus of the text research. Because the data source of the prediction problem of click-through rate is often the online real service log, this paper will mine the valid features of annihilation in a large amount of noise through the feature engineering of the system, and make every effort to use the least features to bring the best prediction effect of the model. In the training stage of the model, the logical regression model suitable for discrete high-dimensional feature scenes and the factoring machine model for sparse feature scenes are selected, and the model is fused with the upstream GBDT model by Stacking integration algorithm, respectively. To achieve the purpose of improving the prediction effect of the model. After the initial implementation of the platform, through functional testing and performance testing, found the platform problems. Through further optimization and iteration, the platform is built. Finally, the content of the paper is summarized, and the improvement direction of the platform is prospected.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP311.52
【相似文獻(xiàn)】
相關(guān)期刊論文 前1條
1 ;新浪微博與淘寶合作 推信息流展示廣告[J];互聯(lián)網(wǎng)天地;2013年04期
相關(guān)重要報(bào)紙文章 前7條
1 記者 霍鑫;網(wǎng)絡(luò)展示廣告智能化 搜索引擎巨頭發(fā)力[N];中國高新技術(shù)產(chǎn)業(yè)導(dǎo)報(bào);2011年
2 本報(bào)記者 焦麗莎;谷歌瞄準(zhǔn)展示廣告[N];中國經(jīng)濟(jì)時(shí)報(bào);2012年
3 本報(bào)記者 方方;谷歌發(fā)力展示廣告[N];中國經(jīng)濟(jì)導(dǎo)報(bào);2011年
4 記者 李蕾;Tim Andree：從注意力導(dǎo)向到價(jià)值導(dǎo)向的轉(zhuǎn)變[N];第一財(cái)經(jīng)日?qǐng)?bào);2014年
5 本報(bào)記者 許泳;Google 展示廣告網(wǎng)絡(luò)化繁為簡[N];計(jì)算機(jī)世界;2011年
6 劉q，
本文編號(hào)：2236975
本文鏈接：http://sikaile.net/kejilunwen/ruanjiangongchenglunwen/2236975.html
最近更新
教材專著
