Click-through rate prediction model integrating user interest and multi-head attention mechanism

Abstract

The purpose of click-through rate (CTR) prediction is to anticipate how likely a person is to click on an advertisement or item. It's required for a lot of internet applications, such online advertising and recommendation systems. The previous click-through rate estimation approach suffered from the following two flaws. On the one hand, input characteristics (such as user id, user age, user age, item id, item category) are usually sparse and multidimensional, making them effective. High-level combination characteristics are used for prediction. Obtaining it manually by domain experts takes a long time and is difficult to finish; also, customer interests are not all the same. The accuracy of the model findings will significantly increase if this immediately recognized component is incorporated in the prediction model. As a consequence, this study creates an IARM (interactive attention rate estimation model) that incorporates user interest as well as a multi-head self-attention mechanism. The deep learning network is used in the model to determine the user's interest expression based on user attributes. The multi-head self-attention mechanism with residual network is then employed to get feature interaction, which enhances the degree of effect of significant characteristics on the estimation result as well as its accuracy. The IARM model outperforms other recent prediction models in the assessment metrics AUC and LOSS, and it has superior accuracy, according to the results from the public experimental data set.

GMINN: Gate‐enhanced multi‐space interaction neural networks for click‐through rate prediction

“…Multi‐head self‐attention mechanism has been extensively applied and proven effective in various domains since its introduction. DCIN‐Attention 23 and IARM 24 modeling high‐order feature interaction in CTR prediction using multi‐head self‐attention neural networks. AutoInt 25 leverages the multi‐head self‐attention mechanism and residual connections to explicitly learn feature interactions, enabling the model to fuse feature information from multiple subspaces.…”

Air pressure prediction model based on the fusion of laser-induced plasma images and spectra

“…[47][48][49] Due to its feature extraction capabilities, it can also capture relationships and features at different positions in images, thereby enhancing a model's understanding and expressive capabilities. It is widely used in many elds, [50][51][52] such as image classication. [53][54][55] In this experiment, it was introduced into the Image module to enhance the model's feature extraction capabilities and improve the pressure prediction performance.…”

Adaptive Feature Cross-Compression for Credit Default Prediction

“…The gate filtering layer function to sift and filter input features, conserving critical feature information and inhibiting less significant ones [24]. By learning feature weights, the process highlights valid information before feature interaction, dynamically attributing significance to the original embedded vectors.…”