基于深度神经网络的电信用户喜爱包预测

Tingshun Li, Huiyu Yang, Dadi Wang, Zesan Liu
{"title":"基于深度神经网络的电信用户喜爱包预测","authors":"Tingshun Li, Huiyu Yang, Dadi Wang, Zesan Liu","doi":"10.1109/NetCIT54147.2021.00048","DOIUrl":null,"url":null,"abstract":"With the popularity of mobile phone, telecom companies have launched more and more package services. It is very difficult to choose telecom user package suitable for him or her. This paper provides a method to build a model based on deep neural network (DNN) for multi-classification, which has high accuracy to help user to pick out a favorite one from dozens of packages. The model is trained out from telecom user big data. First, the telecom big data are preprocessed to be completed integrity, normalized, balanced., and then divided into training data set, validating data set and testing data set. Second, feature engineering needs to be done in order to improve prediction model. Two types of feature engineering are presented in this work to compare which is better. Manual feature engineering is a way to build new features from expertise, while auto feature engineering is from third-party library. Third, an experimental process is design out to obtain prediction models, and use them to predict the testing data set. Finally, some conclusions are obtained by analyzing the experimental results. So, this paper proposes a multi-classification model to recommend a suitable package to a telecom user, which is trained out from telecom big data with high accuracy, more practical. Moreover, this work show that feature engineering and data preprocessing are helpful to obtain better machine learning model.","PeriodicalId":378372,"journal":{"name":"2021 International Conference on Networking, Communications and Information Technology (NetCIT)","volume":"127 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Predicting Telecommunication User Favorite Package by Using Deep Neural Network\",\"authors\":\"Tingshun Li, Huiyu Yang, Dadi Wang, Zesan Liu\",\"doi\":\"10.1109/NetCIT54147.2021.00048\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"With the popularity of mobile phone, telecom companies have launched more and more package services. It is very difficult to choose telecom user package suitable for him or her. This paper provides a method to build a model based on deep neural network (DNN) for multi-classification, which has high accuracy to help user to pick out a favorite one from dozens of packages. The model is trained out from telecom user big data. First, the telecom big data are preprocessed to be completed integrity, normalized, balanced., and then divided into training data set, validating data set and testing data set. Second, feature engineering needs to be done in order to improve prediction model. Two types of feature engineering are presented in this work to compare which is better. Manual feature engineering is a way to build new features from expertise, while auto feature engineering is from third-party library. Third, an experimental process is design out to obtain prediction models, and use them to predict the testing data set. Finally, some conclusions are obtained by analyzing the experimental results. So, this paper proposes a multi-classification model to recommend a suitable package to a telecom user, which is trained out from telecom big data with high accuracy, more practical. Moreover, this work show that feature engineering and data preprocessing are helpful to obtain better machine learning model.\",\"PeriodicalId\":378372,\"journal\":{\"name\":\"2021 International Conference on Networking, Communications and Information Technology (NetCIT)\",\"volume\":\"127 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-12-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 International Conference on Networking, Communications and Information Technology (NetCIT)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/NetCIT54147.2021.00048\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Conference on Networking, Communications and Information Technology (NetCIT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/NetCIT54147.2021.00048","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

摘要

随着手机的普及,电信公司推出了越来越多的套餐服务。选择适合自己的电信用户套餐是一件非常困难的事情。本文提出了一种基于深度神经网络(deep neural network, DNN)的多分类模型的构建方法,该模型具有较高的准确率,可以帮助用户从数十个包中挑选出最喜欢的包。该模型是从电信用户大数据中训练出来的。首先,对电信大数据进行预处理,完成完整性、规格化、均衡化。,然后分为训练数据集、验证数据集和测试数据集。其次,为了改进预测模型,需要进行特征工程。本文提出了两种类型的特征工程,比较哪一种更好。手动特征工程是一种从专业知识中构建新特性的方法,而自动特征工程则来自第三方库。第三,设计了一个实验流程来获得预测模型,并利用预测模型对测试数据集进行预测。最后,通过对实验结果的分析,得出了一些结论。为此,本文提出了一种基于电信大数据训练出来的多分类模型,为电信用户推荐合适的套餐,准确率高,更实用。此外,本研究表明特征工程和数据预处理有助于获得更好的机器学习模型。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Predicting Telecommunication User Favorite Package by Using Deep Neural Network
With the popularity of mobile phone, telecom companies have launched more and more package services. It is very difficult to choose telecom user package suitable for him or her. This paper provides a method to build a model based on deep neural network (DNN) for multi-classification, which has high accuracy to help user to pick out a favorite one from dozens of packages. The model is trained out from telecom user big data. First, the telecom big data are preprocessed to be completed integrity, normalized, balanced., and then divided into training data set, validating data set and testing data set. Second, feature engineering needs to be done in order to improve prediction model. Two types of feature engineering are presented in this work to compare which is better. Manual feature engineering is a way to build new features from expertise, while auto feature engineering is from third-party library. Third, an experimental process is design out to obtain prediction models, and use them to predict the testing data set. Finally, some conclusions are obtained by analyzing the experimental results. So, this paper proposes a multi-classification model to recommend a suitable package to a telecom user, which is trained out from telecom big data with high accuracy, more practical. Moreover, this work show that feature engineering and data preprocessing are helpful to obtain better machine learning model.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信