Using Machine Language to Optimize Recommender Systems in Digital Marketing

Abstract

With the dramatic growth of digital data and increasing competition in the online space, recommender systems have emerged as a key tool in improving user experience and increasing conversion rates in digital marketing. In this study, a new machine language-based algorithm called MLBOA has been introduced and implemented. First, large data consisting of 30,000 user interaction samples were collected. Then, numerical features and scores extracted from text comments were generated and preprocessed using simulation. At this stage, normalization operations were performed to standardize the data scale and features extracted from text data (in the form of sentiment scores) were added to numerical data. After data preparation, to train recurrent neural networks, the data were converted into cell arrays so that the function could process them correctly. The model was trained in an initial process with a learning rate of 0.005. Initial results showed that the network's accuracy in classifying test samples was about 68.6 percent on average. In order to improve the model's performance, the MLBOA optimization algorithm was used to automatically adjust the hyperparameters, especially the learning rate. This algorithm identified the optimal learning rate by running several iterations, and after retraining the network with the obtained optimal learning rate, the network's accuracy increased significantly and reached 74.9 percent. And the prediction error decreased to 0.274. Finally, the results obtained showed that the proposed approach, despite challenges such as the need for advanced computing infrastructure, has a significant improvement in recommendation accuracy and data processing speed compared to traditional methods and can be used as an effective tool in improving recommender systems in digital marketing. This article first introduces a new approach and then analyzes the challenges of implementing machine language-based algorithms and suggests solutions to overcome them.