Pith. sign in

REVIEW 5 major objections 6 minor 50 references

LoRMIkA: Local rule-based model interpretability with k-optimal associations

T0 review · 5 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Mining the k best association rules in a local neighbourhood explains a black-box prediction with four rule types.

desk verdict LoRMIkA's four-rule taxonomy and OPUS-based search are a real step forward for local rule-based explainability, but the reported metrics do not yet establish the accuracy claim. read the letter →

arxiv 1908.03840 v2 pith:YTGY3CIB submitted 2019-08-11 cs.LG cs.AIstat.ML

Dilini Rajapaksha , Christoph Bergmeir , Wray Buntine This is my paper · ORCID
classification cs.LGcs.AIstat.ML
keywords localinterpretabilityclassassociationrulesk-optimalOPUSsearchcounterfactualexplanationsmodel-agnosticexplainerrule-basedtabularclassification
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

LoRMIkA is a model-agnostic explainer for tabular classifiers: it builds a local neighbourhood around the instance to be explained, labels it with the black-box model's predictions, and mines class association rules from that local dataset. A class association rule is an if-then statement whose condition is a set of feature values and whose conclusion is the predicted class. The paper's central argument is that the most predictive rules are not always the best explanations, so the rule search is driven by a user-chosen objective — support, coverage, confidence, lift, or leverage — and returns the k best rules under that objective. The rules are organized into four types: conditions that support the current prediction, conditions that contradict it, hypothetical conditions that would strengthen it, and counterfactual conditions that would reverse it. On the Adult, COMPAS, German, and Covertype datasets, LoRMIkA reports coverage, confidence, lift, stability, and rule simplicity that are competitive with or better than two existing rule-based explainers, Anchor and LORE.

What carries the argument

The central object is the k-optimal class association rule set, mined with the OPUS search algorithm. A class association rule is an if-then rule whose antecedent is a boolean condition on feature values and whose consequent is the predicted class; k-optimal means the top k rules under a user-chosen measure — support, coverage, confidence, lift, or leverage — returned without heuristic pruning. OPUS is an admissible unordered search that guarantees the top-k result and uses Fisher's exact test plus an m-estimate correction to avoid spurious low-coverage rules. The supporting machinery is the neighbourhood generator: an exponential similarity kernel over Euclidean distance selects close training points, and crossover (Equation 8) and mutation (Equation 9) synthesize additional points that inherit categorical values from the closest parent. The four-type contingency table then converts the mined rules into explanations for the specific instance.

What would settle it

On a benchmark where the true local decision boundary is known (for example, a piecewise-linear classifier on synthetic data), build LoRMIkA's neighbourhood and compare each mined rule's confidence and lift against actual nearby training points of the same class; if the rules' confidence collapses when evaluated on genuine close points rather than synthetic ones, the neighbourhood-generation assumption fails.

Watch

Extended reading notes

Core claim

The paper claims that an admissible search called OPUS can find, in a locally generated neighbourhood, the k-optimal class association rules for any of the standard association-rule objectives, and that the resulting rule set gives a complete local explanation. The neighbourhood is formed by selecting the most similar training instances to the explained instance under an exponential similarity kernel, then generating synthetic neighbours by crossover and mutation; all of these instances are labelled by the global model, and rules are mined from the labelled set. Rules are then partitioned according to whether their antecedent matches the instance and whether their consequent matches the prediction, producing four types. The claim is that this yields both predictive rules (high confidence) and interesting rules (high lift or leverage), and that on the tested datasets LoRMIkA matches or beats Anchor and LORE on coverage, confidence, lift, stability, and simplicity, while additionally providing current-contradicting and hypothetical-supporting rules that the baselines do not provide.

Load-bearing premise

The synthetic neighbours generated by crossover and mutation are assumed to carry the same decision logic as the instance being explained, even when they lie off the training-data manifold.

Editorial extensions

If this is right

  • Users can choose the search objective, so the same framework produces predictive explanations (optimized for confidence) or interesting ones (optimized for lift or leverage).
  • The four rule types give a practitioner both the reasons behind a prediction and concrete conditions that would reverse it, without a separate counterfactual search.
  • Because OPUS is admissible, the returned k rules are guaranteed optimal for the chosen objective rather than being a heuristic approximation.
  • In the reported experiments, LoRMIkA achieves higher or comparable coverage, confidence, and lift than Anchor and LORE, with perfect Jaccard stability of 1.00 for the current-supporting rule across runs.
  • LoRMIkA's explanation time stays roughly constant around 38 seconds across global models on Covertype, while the baselines vary more widely.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the objective is a parameter, one could extend LoRMIkA to a multi-objective setting that requires both high coverage and high counterfactual actionability; the paper does not test such a combined objective.
  • The four-type taxonomy suggests a completeness metric that the paper does not define: every true nearby point of the predicted class should be covered by a current supporting rule, and every true nearby point of the opposite class should be covered by a counterfactual rule.
  • The fixed Jaccard stability of 1.00 likely reflects a deterministic neighbourhood selection given the training set; if so, it measures variance in the baselines' sampling rather than a difference in explanatory content.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. The paper proposes LoRMIkA, a model-agnostic local explainer for tabular classifiers. The method selects a neighbourhood of the instance to be explained via Euclidean distance and an exponential similarity kernel, augments it with synthetic instances generated by crossover and mutation (Equations 8 and 9), obtains black-box predictions for the combined set, and mines k-optimal class association rules using the OPUS search algorithm under objectives such as confidence, lift, leverage, coverage, and support. The mined rules are classified into four types — current supporting, current contradicting, hypothetically supporting, and hypothetically contradicting — according to whether their left- and right-hand sides match the explained instance and its prediction. Experiments compare LoRMIkA with Anchor and LORE on COMPAS, Adult, German, and Covertype datasets using coverage, confidence, lift, running time, stability, and rule simplicity, supplemented by a qualitative COMPAS example.

Significance. If the empirical claims were supported, the contribution would be useful: the four-rule typology extends the outputs of Anchor and LORE, and the use of OPUS provides a principled way to optimize different rule-quality measures and to generate both predictive and interesting rules. The high-level algorithm is clearly presented, a GitHub implementation is referenced, and the qualitative COMPAS example is coherent and illustrative. The main value is conceptual and methodological. However, the quantitative evidence for the central claim of competitive local accuracy is currently missing, and the stability result as reported is not credible; these issues prevent the paper from being accepted in its present form.

major comments (5)
  1. [§4.6.1, Tables 1–3, Algorithm 1] The coverage, confidence, and lift values in Tables 1–3 are computed on the same combined neighbourhood CombInst = SelecInst ∪ GenInst that is used to mine the rules in MOGenRules (Algorithm 1, lines 5–7). These are therefore in-sample descriptive statistics of the rules on their own training distribution, not estimates of how well the rules match the black-box model on independent local points. The abstract's claim of 'competitive results in terms of local accuracy' is not tested by these tables. Please add a direct fidelity evaluation: hold out a random subset of the generated neighbourhood, or draw a fresh local sample, and report the precision and coverage of the top-k rules' predictions against the black-box labels on that held-out set.
  2. [§4.6.2, Table 5] The reported Jaccard stability is 1.00 with standard deviation 0.00 for every dataset and every black-box model, including Covertype, despite the randomised neighbourhood sampling in Algorithm 2 (line 19) and the randomised instance generation in Algorithm 3 (α, σ, and random selection of x, y, z). As reported, this is only plausible if the compared runs are not actually independent or if the comparison selects identical rules by construction; otherwise the stability claim is vacuous. Please specify how the 10 runs differ (random seeds, resampling of the 50 instances), report the distribution of Jaccard values rather than only the mean, and state whether the comparison uses the single best rule per objective or the full rule set.
  3. [§3.5, Algorithm 3] Algorithm 3 as written does not implement the parameter NGenInst: it draws one pair x, y for crossover and one triple x, y, z for mutation, and returns a single Icrossover and a single Imutation. There is no loop or repetition that would produce the NGenInst synthetic instances used in Algorithm 1, and the text's statement that '50% of the total instances' are generated by crossover is not reflected in the pseudocode. In the same vein, Section 4.4 does not report the values of k, NGenInst, M, or the number of discretization bins used in the experiments. These omissions make the experimental results unreproducible from the manuscript. Please correct the pseudocode and report the parameter values.
  4. [§3.5, Equations (8)–(9)] The mutation operator Imutation = x + (y − z)·σ with σ ∈ [0.5, 1] can place generated points outside the convex hull of the selected neighbourhood and outside the training data manifold, since y and z are arbitrary neighbours. Section 3.5 simply assumes that black-box predictions on such off-manifold points are informative about the local decision logic of the instance to be explained ('we assume that the generated instance behaves similar...'). This assumption is load-bearing because every rule quality measure in Tables 1–3 is computed on predictions of these generated points. Please provide evidence for this assumption, for example by measuring the distance of generated points to the training data and showing that rules mined with and without mutation points are stable, or by replacing mutation with an in-manifold augmentation procedure.
  5. [§3.7, ℜ−hβ rules] The interpretation of hypothetically contradicting rules as counterfactuals is an overclaim. A rule with LHS=false for the explained instance and RHS=false is a statistical association in the neighbourhood; it does not imply that changing the instance's features to satisfy the LHS would invert the prediction, because (i) confidence is not an interventional estimate and (ii) satisfying the LHS may require changing features jointly in ways that violate feature dependencies and the black-box model's behaviour away from the neighbourhood. The qualitative example ('if this prisoner was older than 29 and priors count was less than 4 ... would be identified as Medium Low') should be rephrased as a description of similar instances that satisfy the LHS, or supported by a proper counterfactual evaluation (e.g., perturbation experiments on the black-box model).
minor comments (6)
  1. [Abstract and §4.1] The abstract says 'three different datasets' and Section 4.1 says 'three real-world datasets', but the experiments actually use four datasets (Adult, COMPAS, German, and Covertype); please correct the inconsistent counts.
  2. [Equation (7)] Equation (7) writes 'x−x′2' where a squared Euclidean norm ||x − x′||² is intended; the notation should be fixed.
  3. [§4.4] Section 4.4 says 'we use a common set of parameters in all three datasets', but four datasets are used; this appears to be a typo.
  4. [Algorithm 2, line 19] Algorithm 2, line 19 uses 'eInstGtrSctGrp.choice(InstM)' without defining InstM or the choice operation; the intended uniform random subsample of size M should be stated explicitly.
  5. [References] References [40] and [41] are both Webb, 'Filtered-top-k association discovery' (2011); the duplicate should be removed and all citations reconciled.
  6. [Table 4] Table 4 reports run-time only for Covertype, and the text says 'average running time over 50 instances'; please clarify whether the same evaluation was performed for the other datasets and why only Covertype is shown.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline quantitative claim is evaluated on the same synthetic neighbourhood used to fit the rules, so the reported coverage, confidence, and lift are in-sample optimization objectives rather than independent tests of local accuracy.

  1. fitted input called prediction [Algorithm 1 lines 5-7 and Section 3.6-3.7, evaluated in Section 4.6.1, Tables 1-3]
    "we obtain the global model predictions of the newly generated instances and the neighbours taken from the training set (i.e., SelecInst∪GenInst) to obtain the global model behaviour for this combined dataset. ..."

    The same combined dataset (SelecInst∪GenInst) is used both to fit the rules and to evaluate them: OPUS selects the top-k rules by confidence, lift, and coverage on that dataset, and Tables 1-3 then report these same three measures on the same neighbourhood. A rule chosen as the highest-confidence rule on the fitting neighbourhood will by construction show high confidence there, so the reported values are in-sample objective values rather than independent estimates of whether the rule matches black-box behavior on held-out local points. The abstract's claim of competitive 'local accuracy' therefore rests on an evaluation that cannot detect rules that are confident on synthetic crossover or mutation points but wrong about the model's actual local behavior.

full rationale

This paper does not rely on a self-citation chain, a uniqueness theorem imported from the authors, or an ansatz hidden in a citation; OPUS is an external algorithm and the neighborhood-generation steps are stated explicitly. The main circularity is in the evaluation: the quantitative claims about coverage, confidence, and lift are computed on exactly the local dataset from which the k-optimal rules were mined, so they are descriptive summaries of the fitting data, not held-out fidelity measurements. This makes the 'local accuracy' claim in the abstract and Section 5 largely untested. A separate red flag is the stability result in Table 5: a Jaccard coefficient of exactly 1.00 with zero variance across randomized neighborhood subsampling and random instance generation is implausible as reported and would need code inspection to determine whether it is an artifact of fixed seeds or a degenerate comparison; however, the paper provides no explicit equation showing that the stability measure is forced by construction, so I do not count it as a demonstrated circular step. Overall, the method itself has independent content, but the central quantitative validation is partially circular, warranting a score of 6.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central method rests on several hand-chosen parameters and domain assumptions: the neighbourhood is defined by a kernel width heuristic, class-balancing caps, an unreported k, and the premise that interpolated or mutated points and inherited categorical values preserve the black-box model's local logic. No new physical entities are introduced.

free parameters (6)
  • Kernel width w = 0.75 * sqrt(number of features)
    Section 3.4, Equation 7: kernel width controls the similarity decay and therefore which training instances enter the neighbourhood. It is set following LIME, not fit to data.
  • L (minimum neighbours per class) = 40
    Algorithm 2 parameter set in Section 4.4 based on preliminary experiments; controls the similarity cut point Sct and neighbourhood size.
  • M or class-balance cap = 1:5 minority-to-majority target proportion
    Section 4.4: cap on the number of neighbours sampled from majority classes, intended to address class imbalance; changes the local training composition.
  • k (number of optimal rules per objective) = not reported
    Section 3.7 says k is user-defined and experiments use k-optimal rules, but the parameter setup in Section 4.4 omits the value of k, making the reported rule sets hard to reproduce exactly.
  • NGenInst (number of synthetic instances) = not reported
    Algorithm 1 takes NGenInst as input and Section 3.5 says 50% are crossover and the rest mutation, but the total number per explained instance is not specified.
  • Discretization bins for continuous features = 3 frequency sub-ranges
    Section 4.2 discretizes continuous attributes into three sub-ranges with approximately equal counts, which determines the granularity of rule antecedents and affects all experiments.
assumptions (5)
  • domain assumption The behaviour of the instance to be explained is similar to the behaviour of instances in its neighbourhood.
    Stated in the Introduction and used in Section 3.4 to justify selecting training instances by Euclidean distance and similarity. This is the standard local-explainability assumption and is not verified per instance.
  • ad hoc to paper Synthetic crossover and mutation points generated by Equations 8 and 9 are valid probes of the black-box model's local decision logic.
    Section 3.5 assumes linear interpolation and mutation of neighbours characterise the local model even though such points can lie outside the training data manifold.
  • ad hoc to paper Newly generated instances inherit categorical values from the closest parent, preserving the behaviour of the instance to be explained.
    Section 3.5 states this assumption explicitly; no empirical check is provided that inherited categorical values are locally faithful.
  • domain assumption Class association rules mined from the combined neighbourhood are faithful local explanations of the global model's prediction for the instance.
    Sections 3.6 and 3.7 use global model predictions on the neighbourhood as targets and assume the mined rules capture the logic behind f(x).
  • standard math OPUS returns k-optimal rules with respect to the chosen objective and its Fisher exact test controls spurious rules.
    Sections 2.2.4 and 3.7 rely on Webb's OPUS guarantees; the result is cited rather than re-derived.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LoRMIkA: Local rule-based model interpretability with k-optimal associations." pith.science (2026). https://pith.science/paper/YTGY3CIB

@misc{pith2026190803840,
  author       = {Pith},
  title        = {Pith review of: LoRMIkA: Local rule-based model interpretability with k-optimal associations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YTGY3CIB}},
  note         = {Machine review of arXiv:1908.03840}
}
read the original abstract

As we rely more and more on machine learning models for real-life decision-making, being able to understand and trust the predictions becomes ever more important. Local explainer models have recently been introduced to explain the predictions of complex machine learning models at the instance level. In this paper, we propose Local Rule-based Model Interpretability with k-optimal Associations (LoRMIkA), a novel model-agnostic approach that obtains k-optimal association rules from a neighbourhood of the instance to be explained. Compared with other rule-based approaches in the literature, we argue that the most predictive rules are not necessarily the rules that provide the best explanations. Consequently, the LoRMIkA framework provides a flexible way to obtain predictive and interesting rules. It uses an efficient search algorithm guaranteed to find the k-optimal rules with respect to objectives such as confidence, lift, leverage, coverage, and support. It also provides multiple rules which explain the decision and counterfactual rules, which give indications for potential changes to obtain different outputs for given instances. We compare our approach to other state-of-the-art approaches in local model interpretability on three different datasets and achieve competitive results in terms of local accuracy and interpretability.

Figures

Figures reproduced from arXiv: 1908.03840 by the authors.

Figure 1
Figure 1. An overview diagram of LoRMIkA to generate local rule-based explanations for the global model [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. An approach to determine instances in the neighbourhood [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Diagram that illustrates the four types of rules generated by LoRMIKA. [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 45 canonical work pages

  1. [1]

    Mining association rules between sets of items in large databases

    Rakesh Agrawal, Tomasz Imieli´ nski, and Arun Swami. Mining association rules between sets of items in large databases. SIGMOD Rec., 22(2):207–216, June 1993

    work page 1993

  2. [2]

    ProtoAttend: Attention-Based Prototypical Learning

    Sercan ¨Omer Arik and Tomas Pfister. Attention-based prototypical learning towards interpretable, confident and robust deep neural networks. CoRR, abs/1902.06292, 2019. URL http://arxiv.org/ abs/1902.06292

    work page Pith review arXiv 1902

  3. [3]

    Interpretable regularized class association rules algorithm for classification in a categorical data space

    Mohamed Azmi, George Runger, and Abdelaziz Berrado. Interpretable regularized class association rules algorithm for classification in a categorical data space. Information sciences, 483:313–331, May

    work page

  4. [4]

    Concreteness and abstraction in everyday explanation

    Christos Bechlivanidis, David A Lagnado, Jeffrey C Zemla, and Steven Sloman. Concreteness and abstraction in everyday explanation. Psychonomic bulletin & review , 24(5):1451–1464, October 2017

    work page 2017

  5. [5]

    Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV)

    Kim Been, Wattenberg Martin, Gilmer Justin, Cai Carrie, Wexler James, Viegas Fernanda, and Sayres Rory. Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV). In Proceedings of the 35 th International Conference on Machine Learning , 35, 2018

    work page 2018

  6. [6]

    An interpretable classification rule mining algo- rithm

    Alberto Cano, Amelia Zafra, and Sebasti´ an Ventura. An interpretable classification rule mining algo- rithm. Information Sciences, 240:1–20, August 2013

    work page 2013

  7. [7]

    Finding interesting associations without support pruning

    E Cohen, M Datar, S Fujiwara, A Gionis, P Indyk, R Motwani, J D Ullman, and C Yang. Finding interesting associations without support pruning. IEEE Trans. Knowl. Data Eng., 13(1):64–78, January 2001

    work page 2001

  8. [8]

    Fast effective rule induction

    William W Cohen. Fast effective rule induction. In Proceedings of the Twelfth International Conference on International Conference on Machine Learning , ICML’95, pages 115–123, San Francisco, CA, USA,

    work page

Show all 50 references
  1. [9]

    BigML, 2019

    Justin Donaldson. BigML, 2019. URL https://www.bigml.com

    2019

  2. [10]

    UCI machine learning repository, 2017

    Dheeru Dua and Casey Graff. UCI machine learning repository, 2017. URL http://archive.ics. uci.edu/ml

    2017

  3. [11]

    Using the m-estimate in rule induction.Journal of computing and information technology , 1(1):37–46, March 1993

    Saˇ so Dˇ zeroski, Bojan Cestnik, and Igor Petrovski. Using the m-estimate in rule induction.Journal of computing and information technology , 1(1):37–46, March 1993

    1993

  4. [12]

    Interestingness measures for data mining: A survey

    Liqiang Geng and Howard J Hamilton. Interestingness measures for data mining: A survey. ACM Computing Surveys (CSUR) , 38(3):9, September 2006

    2006

  5. [13]

    Local rule-based explanations of black box decision systems

    Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Dino Pedreschi, Franco Turini, and Fosca Giannotti. Local rule-based explanations of black box decision systems. CoRR, abs/1805.10820, 2018. URL http://arxiv.org/abs/1805.10820. 24

    2018 arXiv

  6. [14]

    A survey of methods for explaining black box models

    Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Franco Turini, Fosca Giannotti, and Dino Pedreschi. A survey of methods for explaining black box models. ACM Comput. Surv. , 51(5):93:1– 93:42, August 2018

    2018

  7. [15]

    UCI machine learning repository: Statlog (german credit data) data set, November

    Hans Hofmann. UCI machine learning repository: Statlog (german credit data) data set, November

  8. [16]

    compas-analysis, June 2017

    Jeff, Larson and Surya, Mattu and Lauren, Kirchner and Julia, Angwin. compas-analysis, June 2017. URL https://github.com/propublica/compas-analysis

    2017

  9. [17]

    Information complexity in bandit subset selection

    Emilie Kaufmann and Shivaram Kalyanakrishnan. Information complexity in bandit subset selection. In Conference on Learning Theory, pages 228–251, June 2013

    2013

  10. [18]

    Interpreting black box predictions using fisher kernels

    Rajiv Khanna, Been Kim, Joydeep Ghosh, and Oluwasanmi Koyejo. Interpreting black box predictions using fisher kernels. October 2018. URL http://arxiv.org/abs/1810.10118

    2018 arXiv

  11. [19]

    Understanding black-box predictions via influence functions

    Pang Wei Koh and Percy Liang. Understanding black-box predictions via influence functions. In Proceedings of the 34th International Conference on Machine Learning - Volume 70 , ICML’17, pages 1885–1894. JMLR.org, August 2017

    2017

  12. [20]

    On interestingness measures of formal concepts.Information Sciences, 442-443:202–219, May 2018

    S O Kuznetsov and T Makhalova. On interestingness measures of formal concepts.Information Sciences, 442-443:202–219, May 2018

    2018

  13. [21]

    Interpretable decision sets: A joint frame- work for description and prediction

    Himabindu Lakkaraju, Stephen H Bach, and Leskovec Jure. Interpretable decision sets: A joint frame- work for description and prediction. KDD, 2016:1675–1684, August 2016

    2016

  14. [22]

    Faithful and customizable explanations of black box models

    Himabindu Lakkaraju, Ece Kamar, Rich Caruana, and Jure Leskovec. Faithful and customizable explanations of black box models. In Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society, 2019

    2019

  15. [23]

    The mythos of model interpretability

    C LiptonZachary. The mythos of model interpretability. Queueing Systems. Theory and Applications , June 2018

    2018

  16. [24]

    A unified approach to interpreting model predictions

    Scott M Lundberg and Su-In Lee. A unified approach to interpreting model predictions. In I Guyon, U V Luxburg, S Bengio, H Wallach, R Fergus, S Vishwanathan, and R Garnett, editors, Advances in Neural Information Processing Systems 30 , pages 4765–4774. Curran Associates, Inc., 2017

    2017

  17. [25]

    NICGAR: A niching genetic algorithm to mine a diverse set of interesting quantitative association rules

    D Mart´ ın, J Alcal´ a-Fdez, A Rosete, and F Herrera. NICGAR: A niching genetic algorithm to mine a diverse set of interesting quantitative association rules. Information Sciences, 355-356:208–228, August 2016

    2016

  18. [26]

    Measuring interpretability in Rule-Based classification systems

    Detlef D Nauck. Measuring interpretability in Rule-Based classification systems. In Proc. of IEEE International Conference on Fuzzy Systems,2002 , 2002

    2002

  19. [27]

    Supervised descriptive rule discovery: A unifying survey of contrast set, emerging pattern and subgroup mining

    Petra Kralj Novak, Nada Lavraˇ c, and Geoffrey I Webb. Supervised descriptive rule discovery: A unifying survey of contrast set, emerging pattern and subgroup mining. J. Mach. Learn. Res., 10(Feb): 377–403, 2009

    2009

  20. [28]

    Exploration of rule-based knowledge bases: A knowledge engineer’s support

    Agnieszka Nowak-Brzezi´ nska and Alicja Wakulicz-Deja. Exploration of rule-based knowledge bases: A knowledge engineer’s support. Information Sciences, 485:301 – 318, 2019

    2019

  21. [29]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Ga¨ el Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. Journal of machine learning research , 12(Oct):2825–2830, 2011

    2011

  22. [30]

    Interpretable multiclass classification by MDL-based rule lists

    Hugo M Proen¸ ca and Matthijs van Leeuwen. Interpretable multiclass classification by MDL-based rule lists. Information Sciences, October 2019

    2019

  23. [31]

    MAGIX: model agnostic globally interpretable explanations

    Nikaash Puri, Piyush Gupta, Pratiksha Agarwal, Sukriti Verma, and Balaji Krishnamurthy. MAGIX: model agnostic globally interpretable explanations. CoRR, abs/1706.07160, 2017. URL http://arxiv. org/abs/1706.07160

    2017 arXiv

  24. [32]

    FOIL: A midterm report

    J R Quinlan and R M Cameron-Jones. FOIL: A midterm report. In Machine Learning: ECML-93 , pages 1–20. Springer Berlin Heidelberg, 1993

    1993

  25. [33]

    Ross Quinlan

    J. Ross Quinlan. Induction of decision trees. Machine learning, 1(1):81–106, 1986

    1986

  26. [34]

    Why should I trust you?: Explaining the Predictions of Any Classifier

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. Why should I trust you?: Explaining the Predictions of Any Classifier. In Proceedings of the International Conference on Knowledge Discovery and Data Mining (KDDM) , pages 1135–1144, 2016

    2016

  27. [35]

    Anchors: High-Precision Model-Agnostic 25 explanations

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. Anchors: High-Precision Model-Agnostic 25 explanations. In Thirty-Second AAAI Conference on Artificial Intelligence , April 2018

    2018

  28. [36]

    UCI machine learning repository: Adult data set, May 1996

    Kohavi Ronny and Becker Barry. UCI machine learning repository: Adult data set, May 1996. URL https://archive.ics.uci.edu/ml/datasets/adult

    1996

  29. [37]

    Differential evolution – a simple and efficient heuristic for global optimization over continuous spaces

    Rainer Storn and Kenneth Price. Differential evolution – a simple and efficient heuristic for global optimization over continuous spaces. Journal of Global Optimization , 11(4):341–359, December 1997

    1997

  30. [38]

    OPUS: An efficient admissible algorithm for unordered search

    Geoffrey I Webb. OPUS: An efficient admissible algorithm for unordered search. The journal of artificial intelligence research, 3(1):431–465, December 1995. ISSN 1076-9757

    1995

  31. [39]

    Discovering significant rules

    Geoffrey I Webb. Discovering significant rules. In Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , KDD ’06, pages 434–443, New York, NY, USA,

  32. [40]

    Filtered-top- k association discovery

    Geoffrey I Webb. Filtered-top- k association discovery. WIREs Data Mining Knowl Discov , 1(3): 183–192, May 2011

    2011

  33. [41]

    Filtered-top-k association discovery

    Geoffrey I Webb. Filtered-top-k association discovery. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery , 1(3):183–192, 2011

    2011

  34. [42]

    Representer point selection for explaining deep neural networks

    Chih-Kuan Yeh, Joon Kim, Ian En-Hsu Yen, and Pradeep K Ravikumar. Representer point selection for explaining deep neural networks. In S Bengio, H Wallach, H Larochelle, K Grauman, N Cesa-Bianchi, and R Garnett, editors, Advances in Neural Information Processing Systems 31 , pa...

    2018

  35. [43]

    CPAR: Classification based on predictive association rules

    X Yin and J Han. CPAR: Classification based on predictive association rules. In Proceedings of the 2003 SIAM International Conference on Data Mining, Proceedings, pages 331–335. Society for Industrial and Applied Mathematics, May 2003

    2003

  36. [44]

    Leveraging content sensitiveness and user trust- worthiness to recommend Fine-Grained privacy settings for social image sharing

    J Yu, Z Kuang, B Zhang, W Zhang, D Lin, and J Fan. Leveraging content sensitiveness and user trust- worthiness to recommend Fine-Grained privacy settings for social image sharing. IEEE Transactions on Information Forensics and Security , 13(5):1317–1332, May 2018

    2018

  37. [45]

    Beyond bilinear: Generalized multimodal factorized High-Order pooling for visual question answering

    Zhou Yu, Jun Yu, Chenchao Xiang, Jianping Fan, and Dacheng Tao. Beyond bilinear: Generalized multimodal factorized High-Order pooling for visual question answering. IEEE transactions on neural networks and learning systems , 29(12):5947–5959, December 2018

    2018

  38. [46]

    Local Deep-Feature alignment for unsupervised dimension reduction

    Jian Zhang, Jun Yu, and Dacheng Tao. Local Deep-Feature alignment for unsupervised dimension reduction. IEEE transactions on image processing: a publication of the IEEE Signal Processing Society , 27(5):2420–2432, 2018. 26

    2018

  39. [1994]

    URL https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data)

  40. [1995]

    Morgan Kaufmann Publishers Inc

  41. [2006]

    ISBN 9781595933393

    ACM. ISBN 9781595933393

  42. [2019]

    doi: 10.1016/j.ins.2019.01.047

    ISSN 0020-0255. doi: 10.1016/j.ins.2019.01.047. URL https://asu.pure.elsevier.com/en/ publications/interpretable-regularized-class-association-rules-algorithm-for-c

    2019 doi

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.