This paper presents LOX, a neuro-symbolic decision support system for personal finance that integrates neural forecasting with symbolic reasoning. LOX combines Long Short-Term Memory (LSTM) networks for market trend prediction with an ontology-based representation of investment knowledge. The core contribution is a unified pipeline that learns patterns from data, optimizes portfolios based on forecasted returns and risk, and explains recommendations using rule-based reasoning grounded in expert logic. By bridging data-driven learning and symbolic AI, LOX advances the transparency, accuracy, and trustworthiness of automated investment advisors.