Mobile edge computing (MEC) makes up for the disadvantages of cloud computing, and has gained a considerable momentum recently. However, the dynamically changing QoS always results in failures of QoS-ware recommendation and composition of MEC services, which significantly negates the advantages of MEC. To address this issue, considering user-related and service-related contextual factors and various MEC services scheduling scenarios, we propose two context-aware QoS prediction schemes for MEC services. The first scheme is designed for the situations when MEC services are scheduled in real-time, which contains two context-aware real-time QoS estimation methods. One can estimate the real-time multi-QoS of MEC services and the other can estimate the real-time fitted QoS of MEC services. The second scheme is designed for the situations when MEC services are scheduled in the future. This scheme includes two context-aware QoS prediction methods. One can predict the multi-QoS of MEC services and the other can predict the fitted QoS of MEC services. Finally, adaptive QoS prediction strategies are developed in the light of characteristics of the proposed methods. According to these strategies, the most appropriate QoS prediction method could be scheduled adaptively. Extensive experiments are conducted to validate our proposed approaches and to demonstrate their performance.
Accurately predicting stock prices remains a challenging task due to the volatile and complex nature of financial markets. In this study, we propose a novel hybrid deep learning framework that integrates a large language model (LLM), a Linear Transformer (LT), and a Convolutional Neural Network (CNN) to enhance stock price prediction using solely historical market data. The framework leverages the LLM as a professional financial analyst to perform daily technical analysis. The technical indicators, including moving averages (MAs), relative strength index (RSI), and Bollinger Bands (BBs), are calculated directly from historical stock data. These indicators are then analyzed by the LLM, generating descriptive textual summaries. The textual summaries are further transformed into vector representations using FinBERT, a pre-trained financial language model, to enhance the dataset with contextual insights. The FinBERT embeddings are integrated with features from two additional branches: the Linear Transformer branch, which captures long-term dependencies in time-series stock data through a linearized self-attention mechanism, and the CNN branch, which extracts spatial features from visual representations of stock chart data. The combined features from these three modalities are then processed by a Feedforward Neural Network (FNN) for final stock price prediction. Experimental results on the S&P 500 dataset demonstrate that the proposed framework significantly improves stock prediction accuracy by effectively capturing temporal, spatial, and contextual dependencies in the data. This multimodal approach highlights the importance of integrating advanced technical analysis with deep learning architectures for enhanced financial forecasting.
5 February 2024 in Information Processing & Management