Prediction of Stock Prices via Time Series and Machine Learning Models

Basaran B., Cılasun S. M.

5th International Conference on Informatics and Software Engineering, IISEC 2026, Ankara, Türkiye, 5 - 06 Şubat 2026, ss.64-69, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/iisec69317.2026.11418398
  • Basıldığı Şehir: Ankara
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.64-69
  • Anahtar Kelimeler: arima, finance, forecasting, lstm, machine learning, rnn, stock price, time series
  • TED Üniversitesi Adresli: Evet

Özet

This study investigates stock market predictions that are demonstrated via traditional and Machine Learning (ML) models to establish a methodological overview to the traders. Two time-frames were employed in the dataset. In each time-frame, the forecasting performances of Apple (AAPL), Goldman Sachs (GS), The Home Depot (HD), and Tesla (TSLA) were investigated using the ARIMA, ANN, CNN, KNN, and LSTM models. Based on outcomes the performance of the ARIMA have been accomplished favorably in which actual and predicted prices have been matched. Around the ML models' performance, the accomplishments were kept track of with error metrics, absolute metrics and relative bias. According to the outcomes; the overall performances of the models have been compared with that best performing LSTM has been outperformed other models and ANN have been provided second best performance.