Paper: arXiv 2509.14401

Authors: Ahad Yaqoob, Syed M. Abdullah

Abstract

The application of deep learning models for stock price forecasting in emerging markets remains underexplored despite their potential to capture complex temporal dependencies. This study develops and evaluates a Long Short-Term Memory (LSTM) network model for predicting the closing prices of ten major stocks across diverse sectors of the Pakistan Stock Exchange (PSX). Utilizing historical OHLCV data and an extensive set of engineered technical indicators, we trained and validated the model on a multi-year dataset. Our results demonstrate strong predictive performance (\(R^2 > 0.87\)) for stocks in stable, high-liquidity sectors such as power generation, cement, and fertilizers. Conversely, stocks characterized by high volatility, low liquidity, or sensitivity to external shocks (e.g., global oil prices) presented significant forecasting challenges. The study provides a replicable framework for LSTM-based forecasting in data-scarce emerging markets and discusses implications for investors and future research.

Complexity vs Empirical Score

  • Math Complexity: 5.0/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper uses standard deep learning architectures (LSTM) with moderate mathematical complexity, but demonstrates high empirical rigor through detailed data preprocessing, feature engineering with explicit code, and backtest-ready trading strategy analysis on real emerging market data.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Forecast PSX Stock Prices<br>using LSTM Networks"] --> B["Data Collection & Preparation<br>OHLCV + Technical Indicators"]
  B --> C["LSTM Model Architecture<br>Input: Lagged Features<br>Output: Next Day Closing Price"]
  C --> D["Training & Validation<br>Multi-year Historical Data"]
  D --> E["Performance Evaluation<br>R², RMSE, MAE Metrics"]
  E --> F{"Outcome"}
  F --> G["High Accuracy (R² > 0.87)<br>Stable, High-Liquidity Sectors<br>(Power, Cement, Fertilizers)"]
  F --> H["Challenging Forecasts<br>High Volatility, Low Liquidity<br>Oil-Price Sensitive Sectors"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a Long Short-Term Memory (LSTM) network model to predict the closing prices of ten major stocks on the Pakistan Stock Exchange.

Source sentence
This study develops and evaluates a Long Short-Term Memory (LSTM) network model for predicting the closing prices of ten major stocks across diverse sectors of the Pakistan Stock Exchange (PSX).

Assumes 3

  • Historical price patterns are indicative of future performance. model
    Source sentence
    The model assumes historical patterns are indicative of future performance, an assumption that fails during black swan events or structural market shifts.
  • The model does not incorporate external variables such as news sentiment, political events, or global commodity price shocks. data
    Source sentence
    the model does not incorporate external variables like news sentiment, political events, or global commodity price shocks, which are significant drivers of price movements in markets like Pakistan
  • Data are split chronologically into 80% training and 20% testing sets, implying no look‑ahead bias. data
    Source sentence
    The dataset was split chronologically into 80% training and 20% testing sets.

Data

  • Historical daily OHLCV data · public sources (web scraped) market-prices public
    multi-year dataset · ten major PSX stocks across sectors
    Source sentence
    Historical daily OHLCV (Open, High, Low, Close, Volume) data was collected via web scraping of public sources.
  • Fundamental financial indicators · databases (manually collected) fundamentals unknown
    ten major PSX stocks
    Source sentence
    This was supplemented with fundamental financial indicators (e.g., earnings, dividends) manually collected from databases.

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • The LSTM model achieved strong predictive performance with R² greater than 0.87 for stocks in stable, high‑liquidity sectors. predictive-accuracy
    Source sentence
    Our results demonstrate strong predictive performance (R2 > 0.87) for stocks in stable, high-liquidity sectors such as power generation, cement, and fertilizers.
  • Forecasting accuracy was significantly lower for stocks with high volatility, low liquidity, or sensitivity to external shocks. negative
    Source sentence
    Conversely, stocks characterized by high volatility, low liquidity, or sensitivity to external shocks (e.g., global oil prices) presented significant forecasting challenges.
  • LSTM networks captured medium‑to‑long‑term trends effectively for most stocks, especially in stable sectors. performance
    Source sentence
    The model successfully captured medium-to-long-term trends for most stocks.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 33 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (16)

  1. Long Short-Term Memory (1997)
  2. Deep Learning with Long Short-Term Memory Networks for Financial Market Predictions (2018)
  3. Predicting Stock Prices in the Pakistan Market Using Machine Learning and Technical Indicators (2024)
  4. Stock Market Trend Prediction Using Deep Learning (2025)
  5. Study on the Pakistan stock market using a new stock crisis prediction method (2022)
  6. Stock Analysis for New Investors Using Machine Learning (2022) — link
  7. Testing Weak Form Efficiency for Indian Stock Markets (2006) — link
  8. Long-run and Short-run Relationship Between Macroeconomic Variables and Stock Prices in Pakistan: The Case of Lahore Stock Exchange (2009) — link
and 8 more
  1. The Behaviour of Stock Returns in an Emerging Market: A Case Study of Pakistan [with Comments] (1993) — link
  2. Comparative Risk Analysis and Price Prediction of Corporate Shares Using Deep Learning Models Like LSTM and Machine Learning Models (2024) — link
  3. A Multi Parameter Forecasting for Stock Time Series Data Using LSTM and Deep Learning Model (2023) — link
  4. Predictive Modeling for the Moroccan Financial Market: A Nonlinear Time Series and Deep Learning Approach (2025) — link
  5. Enhancing Financial Time Series Forecasting: A Comparative Study of Discrete Wavelet Transform and LSTM Models for Selected Global Indices (2025) — link
  6. Financial Time Series Models—Comprehensive Review of Deep Learning Approaches and Practical Recommendations (2023) — link
  7. Efficient Capital Markets: A Review of Theory and Empirical Work (1970) — link
  8. Analysis of Financial Time Series (2010)

0 of 16 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.