本研究以台指選擇權(TXO) 為研究對象，建構雙軌預測框架並檢驗兩項假說。資訊優勢假說(H1) 檢驗：以逐筆高頻資料建構的Model A，其ATM 隱含波動率次日變動(ΔIV) 的預測誤差，是否顯著低於僅使用公開日資料的Model B。公開資料訊號之經濟可辨識性假說(H2) 則檢驗：在計入散戶等級交易成本(手續費、期交稅與滑價)後，由Model B 訊號驅動的ATM Long/Short Straddle 策略，是否相對HAR-RV 與GARCH(1,1) 兩項時間序列基準提供可辨識之風險調整後績效。
Model A 使用CMoney 日盤與夜盤逐筆資料，建構5 分鐘已實現波動率、跳躍變異與微觀結構特徵；其特徵資訊集嚴格限於t 日13:45 前已實現之資料(日盤t 加上一已完成夜盤)。Model B 僅使用TAIFEX 每日公開揭露變數與Yang–Zhang 估計量。預測模型以長短期記憶神經網路(LSTM) 為主，隨機森林為參照演算法，並採Purged K-Fold 結合Embargo 驗證。樣本期間2018–2026，訓練(2018–2023)、驗證(2024) 與樣本外測試(2025-01–2026-04) 完全分離。預測誤差以Newey-West 修正之Diebold–Mariano 檢定比較；策略層假說以配對Sharpe 差之單尾stationary bootstrap 檢定，並以Holm 逐步法校正多重比較。
實證結果顯示：付費高頻資料於預測誤差與方向準確率均未提供統計顯著優勢，且RMSE 點估計方向於兩組同架構比較不一致(H1 未獲統計支持；DM p = 0.129–0.936，策略層p = 0.951–1.000)；此結果不構成兩類資訊等價之證明。公開資料策略之Sharpe點估計為RF Set B = +1.98、LSTM Set B = +2.17，95% 信賴區間均不含零。相對GARCH 之優勢達5% 顯著(pHolm = 0.034–0.046)；相對HAR-RV 為pHolm = 0.056邊際未達(H2 獲部分統計支持)。由於測試期交易數僅8–11 筆，外推應保守。敏感度分析顯示，上述結論於取樣頻率(30 秒至5 分鐘)、模型家族、交易成本(滑價0–3 tick) 與波動環境子期間下均維持不變。就散戶最應追蹤之指標而言，特徵重要性分析指出公開資料軌道之主要訊號來源依序為結算週期之行事曆結構(週五效應、距結算日天數)、以Yang–Zhang 估計量建構之變異數風險溢酬、TAIEX 短期動量與近期報酬波動；高頻軌道則集中於已實現變異數與期限結構斜率，微觀結構類特徵之邊際貢獻有限。惟個別特徵之重要性估計不確定性高，宜以群組層級解讀。本研究結論嚴格界定於樣本期間、模型設定與成本假設之內，不構成任何投資建議。

This thesis studies the Taiwan Stock Exchange Capitalization Weighted Stock Index Options (TXO) through a dual-track forecasting framework with two testable hypotheses. H1 (Information-Advantage Hypothesis) asks whether Model A, which is built on tick-level data, forecasts the next-day change of ATM implied volatility with significantly lower error than Model B, which is built solely on public daily data. H2 (Economic Identifiability of Public-Data Signals) asks whether Model B signals, traded through an ATM Long/Short Straddle net of retail-grade transaction costs (commissions, transaction tax, and slippage),
generate identifiable risk-adjusted performance relative to the HAR-RV and GARCH(1,1) benchmarks.
Model A constructs 5-minute realized volatility, jump variation, and microstructure features from tick-level day-and-night session data. Its information set is strictly limited to data realized before the 13:45 close of day t (day session t plus the last completed night session). Model B uses only publicly disclosed daily variables from TAIFEX together with the Yang–Zhang estimator. Forecasting employs an LSTM as the primary model and a random
forest benchmark, validated under Purged K-Fold with Embargo. The 2018–2026 sample is split into training (2018–2023), validation (2024), and a chronologically separated out-of-sample test (2025-01–2026-04). Forecast errors are compared with Newey-West-corrected Diebold–Mariano tests. Strategy-level hypotheses use one-sided paired Sharpe-difference stationary-bootstrap tests with Holm step-down adjustment.
Empirically, paid high-frequency data provides no statistically significant advantage in predictive error or directional accuracy. RMSE point estimates point in opposite directions across the two architecture-matched pairs (H1 is not statistically supported; DM p = 0.129–0.936; strategy-level p = 0.951–1.000), and this does not constitute evidence of informational equivalence. The public-data strategies deliver Sharpe point estimates of +1.98 (RF Set B) and +2.17 (LSTM Set B), with 95% confidence intervals excluding zero. The advantage over GARCH is significant at the 5% level (pHolm = 0.034–0.046), while that over HAR-RV is marginally insignificant (pHolm = 0.056) (H2 is partially supported; only 8–11 test trades, so extrapolation should be conservative). Sensitivity analyses across sampling frequencies (30 seconds to 5 minutes), model families, transaction-cost assumptions (0–3 ticks of slippage), and volatility-regime subperiods leave these conclusions unchanged. As to which indicators retail investors should track, permutation importance ranks the settlement-cycle calendar structure (Friday effect, days to settlement), the variance risk premium built on the Yang–Zhang estimator, short-horizon TAIEX momentum, and recent return dispersion as the leading signal sources on the public-data track. The high-frequency track concentrates on realized variance and the term-structure slope, with microstructure features contributing marginally; individual importance estimates are, however, subject to substantial uncertainty and should be read at the group level. All conclusions are strictly bounded by the sample period, model specifications, and cost assumptions of this study, and do not constitute investment advice.

Andersen, T. G., Bollerslev, T., & Diebold, F. X. (2007). Roughing it up: Including jump components in the measurement, modeling, and forecasting of return volatility. Review of Economics and Statistics, 89(4), 701–720.
Andersen, T. G., Bollerslev, T., Diebold, F. X., & Labys, P. (2003). Modeling and forecasting realized volatility. Econometrica, 71(2), 579–625.
Barber, B. M., Lee, Y.-T., Liu, Y.-J., & Odean, T. (2009). Just how much do individual investors lose by trading? Review of Financial Studies, 22(2), 609–632.
Barndorff-Nielsen, O. E., Hansen, P. R., Lunde, A., & Shephard, N. (2008). Designing realized kernels to measure the ex post variation of equity prices in the presence of noise. Econometrica, 76(6), 1481–1536.
Bates, D. S. (1996). Jumps and stochastic volatility: Exchange rate processes implicit in Deutsche mark options. Review of Financial Studies, 9(1), 69–107.
Black, F., & Scholes, M. (1973). The pricing of options and corporate liabilities. Journal of Political Economy, 81(3), 637–654.
Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307–327.
Bollerslev, T., Tauchen, G., & Zhou, H. (2009). Expected stock returns and variance risk premia. Review of Financial Studies, 22(11), 4463–4492.
Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
Bucci, A. (2020). Realized volatility forecasting with neural networks. Journal of Financial Econometrics, 18(3), 502–531.
Carr, P., & Wu, L. (2009). Variance risk premiums. Review of Financial Studies, 22(3), 1311–1341.
Chang, C.-C., Hsieh, P.-F., & Lai, H.-N. (2009). Do informed option investors predict stock returns? Evidence from the Taiwan stock exchange. Journal of Banking and Finance, 33(4), 757–764.
Chang, C.-C., Hsieh, P.-F., & Wang, Y.-H. (2010). Information content of options trading volume for future volatility: Evidence from the Taiwan options market. Journal of Banking and Finance, 34(1), 174–183.
Chen, M.-K., Yang, D.-Y., Hsieh, M.-H., & Wu, M.-E. (2024). An intelligent option trading system based on heatmap analysis via PON/POD yields. Expert Systems with Applications, 257, 124948. https://doi.org/10.1016/j.eswa.2024.124948
Cheng, L.-C., Huang, Y.-H., Hsieh, M.-H., & Wu, M.-E. (2021). A novel trading strategy framework based on reinforcement deep learning for financial market predictions. Mathematics, 9(23), 3094.
Christoffersen, P., Heston, S., & Jacobs, K. (2009). The shape and term structure of the index option smirk: Why multifactor stochastic volatility models work so well. Management Science, 55(12), 1914–1932.
Cont, R. (2001). Empirical properties of asset returns: Stylized facts and statistical issues. Quantitative Finance, 1(2), 223–236.
Corsi, F. (2009). A simple approximate long-memory model of realized volatility. Journal of Financial Econometrics, 7(2), 174–196.
Diebold, F. X., & Mariano, R. S. (1995). Comparing predictive accuracy. Journal of Business & Economic Statistics, 13(3), 253–263.
Dumas, B., Fleming, J., & Whaley, R. E. (1998). Implied volatility functions: Empirical tests. Journal of Finance, 53(6), 2059–2106.
Easley, D., López de Prado, M. M., & O’Hara, M. (2012). Flow toxicity and liquidity in a high-frequency world. Review of Financial Studies, 25(5), 1457–1493.
Easley, D., & O’Hara, M. (1992). Time and the process of security price adjustment. Journal of Finance, 47(2), 577–605.
Engle, R. F. (1982). Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica, 50(4), 987–1007.
Eraker, B., Johannes, M., & Polson, N. (2003). The impact of jumps in volatility and returns. Journal of Finance, 58(3), 1269–1300.
Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, 270(2), 654–669.
Frazzini, A., Israel, R., & Moskowitz, T. J. (2018). Trading costs (Working paper) (SSRN Working Paper No. 3229719). AQR Capital Management. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3229719
Garleanu, N., Pedersen, L. H., & Poteshman, A. M. (2009). Demand-based option pricing. Review of Financial Studies, 22(10), 4259–4299.
Gatheral, J. (2004). A parsimonious arbitrage-free implied volatility parameterization with application to the valuation of volatility derivatives [Conference presentation]. https://www.semanticscholar.org/paper/A-parsimonious-arbitrage-free-implied-volatility-to-Gatheral/0ba2e344d167fb68264aae3315a040d0c5d98b64
Goyal, A., & Saretto, A. (2009). Cross-section of option returns and volatility. Journal of Financial Economics, 94(2), 310–326.
Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. Review of Financial Studies, 33(5), 2223–2273.
Guo, J.-H., & Chang, L.-F. (2020). A generalization of option pricing to price-limit markets. Review of Derivatives Research, 23, 145–161. https://doi.org/10.1007/s11147-019-09160-1
Hansen, P. R., & Lunde, A. (2006). Realized variance and market microstructure noise. Journal of Business & Economic Statistics, 24(2), 127–161.
Heston, S. L. (1993). A closed-form solution for options with stochastic volatility with applications to bond and currency options. Review of Financial Studies, 6(2), 327– 343.
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780.
Hull, J., & White, A. (1987). The pricing of options on assets with stochastic volatilities. Journal of Finance, 42(2), 281–300.
Hull, J. C. (2021). Options, futures, and other derivatives (11th). Pearson.
Hutchinson, J. M., Lo, A. W., & Poggio, T. (1994). A nonparametric approach to pricing and hedging derivative securities via learning networks. Journal of Finance, 49(3), 851–889.
J.P. Morgan/Reuters. (1996). Riskmetrics—technical document (4th ed., Technical Document). J.P. Morgan/Reuters. New York.
Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., & Liu, T.-Y. (2017). LightGBM: A highly efficient gradient boosting decision tree. Advances in Neural Information Processing Systems 30, 3149–3157.
Lee, C. M. C., & Ready, M. J. (1991). Inferring trade direction from intraday data. Journal of Finance, 46(2), 733–746.
Liu, L. Y., Patton, A. J., & Sheppard, K. (2015). Does anything beat 5-minute RV? A comparison of realized measures across multiple asset classes. Journal of Econometrics, 187(1), 293–311.
López de Prado, M. (2018). Advances in financial machine learning. Wiley.
Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems 30, 4765–4774.
Merton, R. C. (1973). Theory of rational option pricing. Bell Journal of Economics and Management Science, 4(1), 141–183.
Merton, R. C. (1976). Option pricing when underlying stock returns are discontinuous. Journal of Financial Economics, 3(1–2), 125–144.
Novy-Marx, R., & Velikov, M. (2016). A taxonomy of anomalies and their trading costs. Review of Financial Studies, 29(1), 104–147.
Pan, J., & Poteshman, A. M. (2006). The information in option volume for future stock prices. Review of Financial Studies, 19(3), 871–908.
Politis, D. N., & Romano, J. P. (1994). The stationary bootstrap. Journal of the American Statistical Association, 89(428), 1303–1313.
Rubinstein, M. (1985). Nonparametric tests of alternative option pricing models using all reported trades and quotes on the 30 most active CBOE option classes from August 23, 1976 through August 31, 1978. Journal of Finance, 40(2), 455–480.
Sirignano, J., & Cont, R. (2019). Universal features of price formation in financial markets: Perspectives from deep learning. Quantitative Finance, 19(9), 1449–1459.
Sundararajan, M., Taly, A., & Yan, Q. (2017). Axiomatic attribution for deep networks. Proceedings of the 34th International Conference on Machine Learning, 3319–3328.
Tharp, V. K. (1999). Trade your way to financial freedom. McGraw-Hill.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems 30, 5998–6008.
Whaley, R. E. (2000). The investor fear gauge. Journal of Portfolio Management, 26(3), 12–17.
Yang, D., & Zhang, Q. (2000). Drift-independent volatility estimation based on high, low, open, and close prices. Journal of Business, 73(3), 477–492.
Yang, Y.-H., Tsaih, R.-H., & Hsieh, M.-H. (2011). A systematic design for coping with model risk. Expert Systems with Applications, 38(6), 7380–7386. https://doi.org/10.1016/j.eswa.2010.12.088
Zhang, L., Mykland, P. A., & Aït-Sahalia, Y. (2005). A tale of two time scales: Determining integrated volatility with noisy high-frequency data. Journal of the American Statistical Association, 100(472), 1394–1411.