sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction
For estimating the click-through rate of advertisements, there are some problems in that the features cannot be automatically constructed, or the features built are relatively simple, or the high-order combination features are difficult to learn under sparse data. To solve these problems, we propose...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-02-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/9/2/350 |
| _version_ | 1798040169520037888 |
|---|---|
| author | Baohua Qiang Yongquan Lu Minghao Yang Xianjun Chen Jinlong Chen Yawei Cao |
| author_facet | Baohua Qiang Yongquan Lu Minghao Yang Xianjun Chen Jinlong Chen Yawei Cao |
| author_sort | Baohua Qiang |
| collection | DOAJ |
| description | For estimating the click-through rate of advertisements, there are some problems in that the features cannot be automatically constructed, or the features built are relatively simple, or the high-order combination features are difficult to learn under sparse data. To solve these problems, we propose a novel structure multi-scale stacking pooling (MSSP) to construct multi-scale features based on different receptive fields. The structure stacks multi-scale features bi-directionally from the angles of depth and width by constructing multiple observers with different angles and different fields of view, ensuring the diversity of extracted features. Furthermore, by learning the parameters through factorization, the structure can ensure high-order features being effectively learned in sparse data. We further combine the MSSP with the classical deep neural network (DNN) to form a unified model named sDeepFM. Experimental results on two real-world datasets show that the sDeepFM outperforms state-of-the-art models with respect to area under the curve (AUC) and log loss. |
| first_indexed | 2024-04-11T22:03:46Z |
| format | Article |
| id | doaj.art-00ce20d82db44336a01801d9241481be |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-04-11T22:03:46Z |
| publishDate | 2020-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-00ce20d82db44336a01801d9241481be2022-12-22T04:00:48ZengMDPI AGElectronics2079-92922020-02-019235010.3390/electronics9020350electronics9020350sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate PredictionBaohua Qiang0Yongquan Lu1Minghao Yang2Xianjun Chen3Jinlong Chen4Yawei Cao5Guangxi Cloud Computing and Big Data Collaborative Innovation Center, Guilin 541004, ChinaGuangxi Cloud Computing and Big Data Collaborative Innovation Center, Guilin 541004, ChinaGuangxi Key Laboratory of Image Graphics and Intelligent Processing, Guilin University of Electronic Technology, Guilin 541004, ChinaGuangxi Key Laboratory of Image Graphics and Intelligent Processing, Guilin University of Electronic Technology, Guilin 541004, ChinaGuangxi Key Laboratory of Image Graphics and Intelligent Processing, Guilin University of Electronic Technology, Guilin 541004, ChinaGuangxi Cloud Computing and Big Data Collaborative Innovation Center, Guilin 541004, ChinaFor estimating the click-through rate of advertisements, there are some problems in that the features cannot be automatically constructed, or the features built are relatively simple, or the high-order combination features are difficult to learn under sparse data. To solve these problems, we propose a novel structure multi-scale stacking pooling (MSSP) to construct multi-scale features based on different receptive fields. The structure stacks multi-scale features bi-directionally from the angles of depth and width by constructing multiple observers with different angles and different fields of view, ensuring the diversity of extracted features. Furthermore, by learning the parameters through factorization, the structure can ensure high-order features being effectively learned in sparse data. We further combine the MSSP with the classical deep neural network (DNN) to form a unified model named sDeepFM. Experimental results on two real-world datasets show that the sDeepFM outperforms state-of-the-art models with respect to area under the curve (AUC) and log loss.https://www.mdpi.com/2079-9292/9/2/350neural networksdeep learningfeatures constructionrecommendationclick-through prediction |
| spellingShingle | Baohua Qiang Yongquan Lu Minghao Yang Xianjun Chen Jinlong Chen Yawei Cao sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction Electronics neural networks deep learning features construction recommendation click-through prediction |
| title | sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction |
| title_full | sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction |
| title_fullStr | sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction |
| title_full_unstemmed | sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction |
| title_short | sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction |
| title_sort | sdeepfm multi scale stacking feature interactions for click through rate prediction |
| topic | neural networks deep learning features construction recommendation click-through prediction |
| url | https://www.mdpi.com/2079-9292/9/2/350 |
| work_keys_str_mv | AT baohuaqiang sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction AT yongquanlu sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction AT minghaoyang sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction AT xianjunchen sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction AT jinlongchen sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction AT yaweicao sdeepfmmultiscalestackingfeatureinteractionsforclickthroughrateprediction |