心理科学进展 ›› 2024, Vol. 32 ›› Issue (10): 1736-1756.doi: 10.3724/SP.J.1042.2024.01736
• 研究方法 • 上一篇
郭鸣谦1, 潘晚坷2, 胡传鹏2
收稿日期:
2023-06-25
出版日期:
2024-10-15
发布日期:
2024-08-13
通讯作者:
郭鸣谦, E-mail: mqguo30@gmail.com; 胡传鹏, E-mail: hcp4715@hotmail.com
GUO Mingqian, PAN Wanke, HU Chuanpeng
Received:
2023-06-25
Online:
2024-10-15
Published:
2024-08-13
摘要: 认知建模近年来在科学心理学获得广泛应用, 而模型比较是认知建模中关键的一环: 研究者需要通过模型比较选择出最优模型, 才能进行后续的假设检验或潜变量推断。模型比较不仅要考虑模型对数据的拟合(平衡过拟合与欠拟合), 也需要考虑模型的复杂度。然而, 模型比较指标众多, 纷繁复杂, 给研究者的选用带来困难。本文将认知建模常用的模型比较指标分为三大类并介绍其计算方法及优劣, 包括拟合优度指标(包括均方误差、决定系数、ROC曲线等)、基于交叉验证的指标(包括AIC、DIC等)和基于边际似然的指标。结合正交Go/No-Go范式的公开数据, 本文展示各指标在R语言中如何实现。在此基础上, 本文探讨各指标的适用情境及模型平均等新思路。
郭鸣谦, 潘晚坷, 胡传鹏. (2024). 认知建模中模型比较的方法. 心理科学进展 , 32(10), 1736-1756.
GUO Mingqian, PAN Wanke, HU Chuanpeng. (2024). Model comparison in cognitive modeling. Advances in Psychological Science, 32(10), 1736-1756.
[1] 胡传鹏, 孔祥祯, Wagenmakers, E.-J., Ly A.,彭凯平. (2018). 贝叶斯因子及其在JASP中的实现. [2] 区健新, 吴寅, 刘金婷, 李红.(2020). 计算精神病学: 抑郁症研究和临床应用的新视角. [3] 王允宏, van den Berg, D., Aust F., Ly A., Wagenmaker E.-J., 胡传鹏. (2023). 贝叶斯方差分析在JASP中的实现. 心理技术与应用, 11(9), 528-541. http://dx.doi.org/10.16842/j.cnki.issn2095-5588.2023.09.002 [4] Acerbi L., Dokka K., Angelaki D. E., & Ma W. J. (2018). Bayesian comparison of explicit and implicit causal inference strategies in multisensory heading perception. [5] Ahn W. Y., Haines N., & Zhang L. (2017). Revealing neurocomputational mechanisms of reinforcement learning and decision-making with the hBayesDM package. [6] Akaike, H. (1974). A new look at the statistical model identification. [7] Anderson, D., & Burnham, K. (2004). [8] Ballard I. C., Wagner A. D., & McClure S. M. (2019). Hippocampal pattern separation supports reinforcement learning. [9] Betts, M. J., Richter, A., de Boer, L., Tegelbeckers, J., Perosa, V., Baumann, V.,. Krauel, K.(2020). Learning in anticipation of reward and punishment: Perspectives across the human lifespan. [10] Bishop, C. M. (2006). [11] Blei D. M., Kucukelbir A.,& McAuliffe, J. D.(2017). Variational inference: A review for statisticians. [12] Boehm U., Annis J., Frank M. J., Hawkins G. E., Heathcote A., Kellen D.,. Wagenmakers, E.-J.(2018). Estimating across-trial variability parameters of the Diffusion Decision Model: Expert advice and recommendations. [13] Boehm U., Evans N. J., Gronau Q. F., Matzke D., Wagenmakers E.-J., & Heathcote, A. J. (2023). Inclusion Bayes factors for mixed hierarchical diffusion decision models. Psychological Methods. https://doi.org/10.1037/met0000582Bos, C. S. (2002). A comparison of marginal likelihood computation methods. In: Härdle, W., & Rönz, B. (Eds.), Compstat: Physica, Heidelberg. [14] Brown S. D.,& Heathcote, A.(2008). The simplest complete model of choice response time: linear ballistic accumulation. [15] Burnham, K. P., & Anderson, D. R. (2004). Multimodel inference: Understanding AIC and BIC in model selection. [16] Carpenter B., Gelman A., Hoffman M. D., Lee D., Goodrich B., Betancourt M.,. Riddell A. (2017). Stan: A probabilistic programming language. [17] Casella, G., & Berger, R. L. (2002). [18] Cavanagh J. F., Eisenberg I., Guitart-Masip M., Huys Q., & Frank M. J. (2013). Frontal theta overrides pavlovian learning biases. [19] Clyde M. A., Ghosh J.,& Littman, M. L.(2011). Bayesian adaptive sampling for variable selection and model averaging. [20] Collins, A. G., & Frank, M. J. (2012). How much of reinforcement learning is working memory, not reinforcement learning? A behavioral, computational, and neurogenetic analysis. [21] Collins, A. G. E., & Frank, M. J. (2018). Within- and across- trial dynamics of human EEG reveal cooperative interplay between reinforcement learning and working memory. [22] Daniel R., Radulescu A., & Niv Y. (2020). Intact reinforcement learning but impaired attentional control during multidimensional probabilistic learning in older adults. [23] Daunizeau J., Adam V., & Rigoux L. (2014). VBA: A probabilistic treatment of nonlinear models for neurobiological and behavioural data. [24] Davis, J., & Goadrich, M. (2006). [25] Daw N. D.(2011). Trial-by-trial data analysis using computational models. In Delgado, M. R. (Ed.), Decision making, affect, learning: Attention performance XXIII (Vol. 23, pp. 3-38). Oxford University Press. [26] Dayan P., Niv Y., Seymour B.,& Daw, N. D.(2006). The misbehavior of value and the discipline of the will. [27] Devine S., Falk C. F., & Fujimoto K. A. (2023). Comparing the accuracy of three predictive information criteria for Bayesian linear multilevel model selection. [28] Dickey, J. (1973). Scientific reporting and personal probabilities: Student's hypothesis. [29] Dickey, J. M. (1976). Approximate posterior distributions. [30] Donkin C., Heathcote A., & Brown S. (2009). [31] Dorfman, H. M., & Gershman, S. J. (2019). Controllability governs the balance between Pavlovian and instrumental action selection. [32] Doucet A.,& Johansen, A. M. (2009). A tutorial on particle filtering and smoothing: Fifteen years later In Crisan, D (Ed), Handbook of nonlinear filtering Oxford University Press Fifteen years later. [33] Dziak J. J., Coffman D. L., Lanza S. T., Li R., & Jermiin L. S. (2020). Sensitivity and specificity of information criteria. [34] Evans, N. J. (2019). Assessing the practical differences between model selection methods in inferences about choice response time tasks. [35] Evans N. J., Hawkins G. E., & Brown S. D. (2020). The role of passing time in decision-making. [36] Farrell S.,& Lewandowsky, S. (2018). Computational modeling of cognition and behavior. Cambridge University Press. [37] Fong, E., & Holmes, C. C. (2020). On the marginal likelihood and cross-validation. [38] Fontanesi L., Gluth S., Spektor M. S., & Rieskamp J. (2019). A reinforcement learning diffusion decision model for value-based decisions. [39] Forstmann B. U., Ratcliff R., & Wagenmakers E. J. (2016). Sequential sampling models in cognitive neuroscience: Advantages, applications, and extensions. [40] Friedman J., Hastie T., & Tibshirani R. (2001). [41] Friston K., Kilner J.,& Harrison, L.(2006). A free energy principle for the brain. [42] Friston K., Mattout J.,Trujillo-Barreto, N., Ashburner, J., & Penny, W.(2007). Variational free energy and the Laplace approximation. [43] Gamerman, D., & Lopes, H. F. (2006). [44] Geisser S.,& Eddy, W. F.(1979). A predictive approach to model selection. [45] Gelfand, A. E., & Dey, D. K. (1994). Bayesian model choice: Asymptotics and exact calculations. [46] Gelman A., Carlin J. B., Stern H. S., Dunson D. B., Vehtari A., & Rubin D. B. (2013). [47] Gelman A., Hwang J., & Vehtari A. (2013). Understanding predictive information criteria for Bayesian models. [48] Gershman S. J.(2016). Empirical priors for reinforcement learning models. [49] Gronau Q. F., Sarafoglou A., Matzke D., Ly A., Boehm U., Marsman M.,. Steingroever, H.(2017). A tutorial on bridge sampling. [50] Gronau, Q. F., & Wagenmakers, E. J. (2019). Limitations of Bayesian Leave-One-Out Cross-Validation for model selection. Computational Brain & Behavior, 2(1), 1-11. https://doi.org/10.1007/s42113-018-0011-7 [51] Guitart-Masip,M., Huys, Q. J., Fuentemilla, L., Dayan, P., Duzel, E., & Dolan, R. J.(2012). Go and no-go learning in reward and punishment: Interactions between affect and effect. [52] Hair J. F., Black W. C., Babin B. J., & Anderson R. E. (2010). [53] Hammersley, J. (2013). [54] Heck, D. W. (2019). A caveat on the Savage-Dickey density ratio: The case of computing Bayes factors for regression parameters. [55] Hinne M., Gronau Q. F., van den Bergh D., & Wagenmakers E.-J. (2020). A conceptual introduction to Bayesian model averaging. [56] Hurvich, C. M., & Tsai, C.-L. (1989). Regression and time series model selection in small samples. [57] Huys Q. J., Cools R., Gölzer M., Friedel E., Heinz A., Dolan R. J., & Dayan P. (2011). Disentangling the roles of approach, activation and valence in instrumental and pavlovian responding. [58] Huys Q. J., Maia T. V., & Frank M. J. (2016). Computational psychiatry as a bridge from neuroscience to clinical applications. [59] Iglesias S., Mathys C., Brodersen K. H., Kasper L., Piccirelli M., den Ouden, H. E., & Stephan, K. E.(2013). Hierarchical prediction errors in midbrain and basal forebrain during sensory learning. [60] Ikink, I., Engelmann, J. B., van den Bos, W., Roelofs, K., & Figner, B.(2019). Time ambiguity during intertemporal decision-making is aversive, impacting choice and neural value coding. [61] Kass R. E.,& Raftery, A. E.(1995). Bayes factors. [62] Körding K. P.,& Wolpert, D. M.(2006). Bayesian decision theory in sensorimotor control. [63] Kvålseth T. O.(1985). Cautionary note about R2. [64] Lebreton M., Bacily K., Palminteri S., & Engelmann J. B. (2019). Contextual influence on confidence judgments in human reinforcement learning. [65] Li J., Schiller D., Schoenbaum G., Phelps E. A., & Daw N. D. (2011). Differential roles of human striatum and amygdala in associative learning. [66] Li J. A., Dong D., Wei Z., Liu Y., Pan Y., Nori F., & Zhang X. (2020). Quantum reinforcement learning during human decision-making. [67] Li, Z.-W., & Ma, W. J. (2021). An uncertainty-based model of the effects of fixation on choice. [68] MacKay, D. J. (2003). Information theory, inference and learning algorithms. Cambridge University Press. [69] McFadden, D. L. (1984). Chapter 24 Econometric analysis of qualitative response models. In Durlauf, S. N. (Ed.), [70] Menard, S. (2000). Coefficients of determination for multiple logistic regression analysis. [71] Meng, X.-L., & Wong, W. H. (1996). Simulating ratios of normalizing constants via a simple identity: A theoretical exploration. [72] Merlise, C., & Edward, I. G. (2004). Model uncertainty. [73] Montague P. R., Dolan R. J., Friston K. J.,& Dayan, P.(2012). Computational psychiatry. [74] Murphy K. P.(2023). Probabilistic machine learning: An introduction. The MIT Press. [75] Myung, I. J., & Pitt, M. A. (1997). Applying Occam’s razor in modeling cognition: A Bayesian approach. [76] Myung, J., & Pitt, M. (2018). Model comparison in psychology. In Wagenmakers, E.J. (Ed.), [77] Palminteri S., Wyart V.,& Koechlin, E.(2017). The importance of falsification in computational cognitive modeling. [78] Pedersen M. L., Ironside M., Amemori K. I., McGrath C. L., Kang M. S., Graybiel A. M., Pizzagalli D. A., & Frank M. J. (2021). Computational phenotyping of brain- behavior dynamics underlying approach-avoidance conflict in major depressive disorder. [79] Raab, H. A., & Hartley, C. A. (2020). Adolescents exhibit reduced Pavlovian biases on instrumental learning. Scientific reports, 10(1), 15770. https://doi.org/10.1038/s41598-020- 72628-w [80] Ratcliff R., Smith P. L., Brown S. D.,& McKoon, G.(2016). Diffusion decision model: Current issues and history. [81] Schultz W., Dayan P., & Montague P. R. (1997). A neural substrate of prediction and reward. [82] Schwarz, G. (1978). Estimating the dimension of a model. [83] Sclove, S. L. (1987). Application of model-selection criteria to some problems in multivariate analysis. [84] Sivula T., Magnusson M., Matamoros A. A.,& Vehtari, A.(2020). Uncertainty in Bayesian leave-one-out cross- validation based model comparison. [85] Spiegelhalter D. J., Best N. G., Carlin B. P., & Van Der Linde, A. (2002). Bayesian measures of model complexity and fit. [86] Spiegelhalter D. J., Best N. G., Carlin B. P., & Van Der Linde, A. (2014). The deviance information criterion: 12 years on. Journal of the Royal Statistical Society. Series B (Statistical Methodology), 76(3), 485-493. http://www. jstor.org/stable/24774528 [87] Steinberg E. E., Keiflin R., Boivin J. R., Witten I. B., Deisseroth K., & Janak P. H. (2013). A causal link between prediction errors, dopamine neurons and learning. [88] Steingroever H., Wetzels R., & Wagenmakers E.-J. (2014). Absolute performance of reinforcement-learning models for the Iowa Gambling Task. [89] Steingroever H., Wetzels R., & Wagenmakers E.-J. (2016). Bayes factors for reinforcement-learning models of the Iowa gambling task. [90] Stephan K. E., Penny W. D., Daunizeau J., Moran R. J.,& Friston, K. J.(2009). Bayesian model selection for group studies. [91] Stone, M. (1977). An asymptotic equivalence of choice of model by cross-validation and Akaike's criterion. [92] Sugiura, N. (1978). Further analysis of the data by akaike's information criterion and the finite corrections: Further analysis of the data by akaike's. [93] Suzuki S., Harasawa N., Ueno K., Gardner J. L., Ichinohe N., Haruno M., Cheng K.,& Nakahara, H.(2012). Learning to simulate others' decisions. [94] Swart J. C., Fröbose M. I., Cook J. L., Geurts D. E., Frank M. J., Cools R., & den Ouden, H. E. (2017). Catecholaminergic challenge uncovers distinct Pavlovian and instrumental mechanisms of motivated (in)action. [95] Vandekerckhove J., Matzke D., & Wagenmakers, E.-J. (2015). Model comparison and the principle of parsimony. In Busemeyer, J. R., Wang, Z., Townsend, J. T., & Eidels, A. (Eds.), The Oxford handbook of computational and mathematical psychology. (pp. 300-319). Oxford University Press. [96] Vandekerckhove J., Tuerlinckx F., & Lee M. D. (2011). Hierarchical diffusion models for two-choice response times. [97] van de Schoot R., Depaoli S., King R., Kramer B., Märtens K., Tadesse M. G.,. Yau C. (2021). Bayesian statistics and modelling. [98] Vehtari, A. (2022). [99] Vehtari A., Gelman A., & Gabry J. (2017). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. [100] Vehtari A., Mononen T., Tolvanen V., Sivula T., & Winther O. (2016). Bayesian leave-one-out cross-validation approximations for Gaussian latent variable models. The Journal of Machine Learning Research, 17(1), 3581-3618. 2016). Bayesian leave-one-out cross-validation approximations for Gaussian latent variable models. The Journal of Machine Learning Research, 17(1), 3581-3618. http://jmlr.org/papers/v17/14-540.html [101] Vehtari A., Simpson D. P., Yao Y., & Gelman A. (2019). Limitations of “Limitations of Bayesian Leave-one-out Cross-Validation for Model Selection”. Computational Brain & Behavior, 2(1), 22-27. https://doi.org/10.1007/s42113-018-0020-6Verstynen, T., & Kording, K. P. [102] Vrieze, S. I. (2012). Model selection and psychological theory: A discussion of the differences between the Akaike information criterion (AIC) and the Bayesian information criterion (BIC). [103] Wagenmakers, E.-J., & Farrell, S. (2004). AIC model selection using Akaike weights. [104] Wagenmakers E. J., Lodewyckx T., Kuriyal H.,& Grasman, R.(2010). Bayesian hypothesis testing for psychologists: A tutorial on the Savage-Dickey method. [105] Wasserman, L. (2006). [106] Watanabe, S. (2010). Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory. Journal of machine learning research, 11(12). http://jmlr.org/papers/v11/watanabe10a. html [107] Westbrook A., van den Bosch R., Määttä J., Hofmans L., Papadopetraki D., Cools R., & Frank M. J. (2020). Dopamine promotes cognitive effort by biasing the benefits versus costs of cognitive work. [108] Wilks, S. S. (1938). The large-sample distribution of the likelihood ratio for testing composite hypotheses. The annals of mathematical statistics, 9(1), 60-62. http://www. jstor.org/stable/2957648 [109] Wilson, R. C., & Collins, A. G. (2019). Ten simple rules for the computational modeling of behavioral data. [110] Yang, Y. (2005). Can the strengths of AIC and BIC be shared? A conflict between model indentification and regression estimation. [111] Yao Y., Vehtari A., Simpson D., & Gelman A. (2018). Using stacking to average Bayesian predictive distributions (with discussion). [112] Zhang L., Lengersdorff L., Mikus N., Gläscher J., & Lamm C. (2020). Using reinforcement learning models in social neuroscience: Frameworks, pitfalls and suggestions of best practices. [113] Zhang Y.,& Yang, Y.(2015). Cross-validation for selecting a model selection procedure. |
[1] | 李兴珊, 张淇玮, 黄林洁琼. 中文词汇语义加工过程的计算模拟与实验验证[J]. 心理科学进展, 2024, 32(9): 1379-1392. |
[2] | 韩宜瑾, 朱金丽, 侯方. 弱视视觉系统的对比度增益研究[J]. 心理科学进展, 2023, 31(suppl.): 4-4. |
[3] | 黎穗卿, 陈新玲, 翟瑜竹, 张怡洁, 章植鑫, 封春亮. 人际互动中社会学习的计算神经机制[J]. 心理科学进展, 2021, 29(4): 677-696. |
[4] | 高青林, 周媛. 计算模型视角下信任形成的心理和神经机制——基于信任博弈中投资者的角度[J]. 心理科学进展, 2021, 29(1): 178-189. |
[5] | 张银花, 李红, 吴寅. 计算模型在道德认知研究中的应用[J]. 心理科学进展, 2020, 28(7): 1042-1055. |
[6] | 李精精, 张剑, 田慧荣, Jeffrey B.Vancouver. 动态计算模型在组织行为学研究中的应用[J]. 心理科学进展, 2020, 28(2): 368-380. |
[7] | 区健新, 吴寅, 刘金婷, 李红. 计算精神病学:抑郁症研究和临床应用的新视角[J]. 心理科学进展, 2020, 28(1): 111-127. |
[8] | 高绍兵, 李永杰. 结合视觉自底向上和自顶向下机制的多光源颜色恒常性算法研究[J]. 心理科学进展, 2019, 27(suppl.): 96-96. |
[9] | 任衍具;傅小兰. 图表加工的理论模型[J]. 心理科学进展, 2006, 14(1): 23-31. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||