Machine Learning-Driven Predictive Modelling for Economic Indicators Using Advanced Data Analytics

This study defines the use of machine learning models to predict car production volume for better planning and correspond them with market desirability. This work assesses three different models: Artificial Neural Network (ANN), Support Vector Machine (SVM) and Long Short-Term Memory (LSTM), using an extensive dataset composed of 10,332 historical production volume records over the past ten years accompanied by raw material prices, energy costs along with labor cost attributes as well as experiments will market demand, seasonal trends economic indicators. The dataset was preprocessed with some techniques like handling of missing data, normalization and also encoding for categorical variables to make it strong. The models were trained and tested using a 70-30 training - test split, evaluated on different performance metrics such as accuracy, precision and other criteria. The best-case accuracy of 97.75% was achieved by the ANN, hence showing its capability to capture complex non-linear relationships within data. The SVM reached 93.40% accuracy, much stronger with high precision and recall tradeoff post update to the split point. The LSTM model providing an accuracy rate of 91.22%, considering it was best suited for handling sequential data. The above results have shown the ANN model better accuracy in predictions and because of this, we can use an ANN for future demand prediction efficiently. This has valuable implications for car manufacturing and is a good direction to elaborate on predictive models in the future.