Application of 6G Driver and Deep Learning in Marketing Data Mining System

The fashion business is shifting towards sustainability, emphasising personalised clothing creation to accommodate various customer tastes while minimising environmental effect. This article proposes using deep learning models in cloud computing infrastructure to produce customised sustainable fashion items. Manual garment design approaches are reliable, ineffective. This study integrates deep learning with sustainability optimisation utilising the Fire hawk Optimisation, for personalised clothing design and categorization. The proposed method includes data pre-processing, feature extraction, model training, final classification using the novel lightweight residual network with Energy Valley Optimizer (LResNet-EVO), and sustainability optimisation to meet the growing demand for tailored fashion solutions while reducing environmental impact. After data cleaning to eliminate unnecessary items and duplicates, dataset variety is increased using augmentation methods such rotation, flipping, scaling, and noise addition. Colour, texture, and form data are extracted using Improved Artificial Neural Networks for personalised design. CNNs are used for personalised clothing creation and categorization, using pre-trained models on the DeepFashion dataset for better performance. LResNet-EVO classifies created designs by style, occasion, or trendiness to improve usability. FA/FHO reduces material waste, optimises manufacturing, and considers end-of-life in sustainability optimisation. Iteratively modifying design parameters to maximise material efficiency and minimise environmental effect throughout the garment lifespan promotes environmentally aware fashion. Integration with cloud infrastructure provides scalability, real-time feedback, and large-scale calculations for personalised fashion solutions. This study makes fashion more accessible, efficient, and ecologically friendly by connecting innovation, sustainability, and scalability.