Abstract
The interpretability of a machine learning model plays a significant role in practical applications, thus it is necessary to develop a method to compare the interpretability for different models so as to select the most appropriate one. However, model interpretability, a highly subjective concept, is difficult to be accurately measured, not to mention the interpretability comparison of different models. To this end, we develop an interpretability evaluation model to compute model interpretability and compare interpretability for different models. Specifically, first we we present a general form of model interpretability. Second, a questionnaire survey system is developed to collect information about users’ understanding of a machine learning model. Next, three structure features are selected to investigate the relationship between interpretability and structural complexity. After this, an interpretability label is build based on the questionnaire survey result and a linear regression model is developed to evaluate the relationship between the structural features and model interpretability. The experiment results demonstrate that our interpretability evaluation model is valid and reliable to evaluate the interpretability of different models.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
Data is available at http://pan.baidu.com/s/1eRSNWtW.
References
Zhou, Z.H.: Machine Learning, 1st edn. Tsinghua University Press, Beijing (2016)
Jovanovic, M., Radovanovic, S., Vukicevic, M., et al.: Building interpretable predictive models for pediatric hospital readmission using Tree-Lasso logistic regression. Artif. Intell. Med. 72, 12 (2016)
Huysmans, J., Dejaeger, K., Mues, C., et al.: An empirical evaluation of the comprehensibility of decision table, tree and rule based predictive models. Decis. Support Syst. 51(1), 141–154 (2011)
Sim, K.C.: On constructing and analysing an interpretable brain model for the DNN based on hidden activity patterns. In: Automatic Speech Recognition and Understanding, pp. 22–29. IEEE (2016)
Allahyari, H., Lavesson, N.: User-oriented assessment of classification model understandability. In: Scandinavian Conference on Artificial Intelligence (2011)
Lipton, Z.C.: The Mythos of Model Interpretability (2016)
Ustun, B., Traca, S., Rudin, C.: Supersparse linear integer models for predictive scoring systems. In: AAAI Conference on Artificial Intelligence (2013)
Lu, T.K.P., Chau, V.T.N., Phung, N.H.: Extracting rule RF in educational data classification: from a random forest to interpretable refined rules. In: International Conference on Advanced Computing and Applications, pp. 20–27. IEEE (2015)
Turner, R.: A Model Explanation System: Latest Updates and Extensions (2016)
Ribeiro, M.T., Singh, S., Guestrin, C.: “Why should i trust you?”: explaining the predictions of any classifier. In: ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 1135–1144. ACM (2016)
Maimon, O., Rokach, L.: Data Mining and Knowledge Discovery Handbook. Springer, Berlin (2005)
Schielzeth, H.: Simple means to improve the interpretability of regression coefficients. Methods Ecol. Evol. 1(2), 103–113 (2010)
Martens, D., Vanthienen, J., Verbeke, W., et al.: Performance of classification models from a user perspective. Decis. Support Syst. 51(4), 782–793 (2011)
Freitas, A.A.: Comprehensible classification models:a position paper. ACM SIGKDD Explor. Newsl. 15(1), 1–10 (2014)
Bibal, A., Frénay, B.: Interpretability of machine learning models and representations: an introduction. In: Proceedings ESANN, pp. 77–82 (2016)
Piltaver, R., Luštrek, M., Gams, M., et al.: What makes classification trees comprehensible? Expert Syst. Appl. 62(10), 333–346 (2016)
Sweller, J.: Cognitive load during problem solving: effects on learning. Cognit. Sci. 12(2), 257–285 (1988)
Piltaver, R., Luštrek, M., Gams, M., et al.: Comprehensibility of classification trees—survey design. In: International Multiconference Information Society—Is (2014)
Xing, W., Guo, R., Petakovic, E., et al.: Participation-based student final performance prediction model through interpretable Genetic Programming. Comput. Hum. Behav. 47(C), 168–181 (2015)
Zhou, L., Si, Y.W., Fujita, H.: Predicting the listing statuses of Chinese-listed companies using decision trees combined with an improved filter feature selection method. Knowl.-Based Syst. 128 (2017)
Gorzałczany, M.B., Rudziński, F.: Interpretable and accurate medical data classification—a multi-objective genetic-fuzzy optimization approach. Expert Syst. Appl. 71, 26–39 (2017)
White box. https://en.wikipedia.org/wiki/White_box_(software_engineering). Accessed 20 Nov 2017
Black box. https://en.wikipedia.org/wiki/Black_box. Accessed 20 Nov 2017
“no free lunch” theorem. https://en.wikipedia.org/wiki/No_free_lunch_theorem. Accessed 19 March 2018
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this paper
Cite this paper
Zhou, Q., Liao, F., Mou, C., Wang, P. (2018). Measuring Interpretability for Different Types of Machine Learning Models. In: Ganji, M., Rashidi, L., Fung, B., Wang, C. (eds) Trends and Applications in Knowledge Discovery and Data Mining. PAKDD 2018. Lecture Notes in Computer Science(), vol 11154. Springer, Cham. https://doi.org/10.1007/978-3-030-04503-6_29
Download citation
DOI: https://doi.org/10.1007/978-3-030-04503-6_29
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-04502-9
Online ISBN: 978-3-030-04503-6
eBook Packages: Computer ScienceComputer Science (R0)