Data-Driven Insights into Automotive Customer Complaints: A Machine Learning Approach to Predictive Analytics

Automotive customer complaints forecast as a difficult-to-solve and data-driven machine learning classification problem. Key findings from background research include the problems associated with data imbalance, shortcomings of the models in terms of temporal pattern analysis and variability of prediction errors. It incorporates clustering and LSTM models to analyze product complaint characteristics, improving accuracy and flexibility. The proposed model operates based on a diverse dataset of automotive complaints, and the model accuracy is 94.5%, the precision is 92.7<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, the recall is 93.1%, the F1 score is 92.9%, and the AUC is 95.3%. Compared to other methods, the advantage of this particular work is that it achieves far better accuracy on complex and imbalanced data sets than previous approaches. The findings proved that the chosen model's approach is rather effective for the prognosis of different kinds of complaints to assist in the usage of the anticipative maintenance approaches. It is useful for automakers, because, knowing that the issues, and indicators of the low degree of customer satisfaction define, manufacturers can come to them before high-risk issues occur.