Contrast Set Mining for Distinguishing Between Similar Diseases

Kralj, Petra; Lavrač, Nada; Gamberger, Dragan; Krstačić, Antonija

doi:10.1007/978-3-540-73599-1_12

Cited by 22 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mining EPs has been well studied since 1999 , and the implementations of EPs in many areas are found in the literature as well. For example, implementations in the medical area can distinguish similar diseases , classify cancer diagnosis data , and profile leukemia patients . In the networking environment, EPs are implemented to find significant differences in network data streams , or to detect masqueraders .…”

Section: Related Workmentioning

confidence: 99%

Novel Push-Front Fibonacci Windows Model for Finding Emerging Patterns with Better Completeness and Accuracy

Akhriza¹,

Ma²,

Li³

2018

ETRI Journal

View full text Add to dashboard Cite

To find the emerging patterns (EPs) in streaming transaction data, the streaming is first divided into some time windows containing a number of transactions. Itemsets are generated from transactions in each window, and then the emergence of itemsets is evaluated between two windows. In the tilted‐time windows model (TTWM), it is assumed that people need support data with finer accuracy from the most recent windows, while accepting coarser accuracy from older windows. Therefore, a limited array's elements are used to maintain all support data in a way that condenses old windows by merging them inside one element. The capacity of elements that accommodates the windows inside is modeled using a particular number sequence. However, in a stream, as new data arrives, the current array updating mechanisms lead to many null elements in the array and cause data incompleteness and inaccuracy problems. Two models derived from TTWM, logarithmic TTWM and Fibonacci windows model, also inherit the same problems. This article proposes a novel push‐front Fibonacci windows model as a solution, and experiments are conducted to demonstrate its superiority in finding more EPs compared to other models.

show abstract

Section: Related Workmentioning

confidence: 99%

Novel Push-Front Fibonacci Windows Model for Finding Emerging Patterns with Better Completeness and Accuracy

Akhriza¹,

Ma²,

Li³

2018

ETRI Journal

View full text Add to dashboard Cite

show abstract

“…However, these methods focus only on maximizing classification accuracy and aim to achieve interpretability just by building the model from rules. Similarly, algorithms for problems such as subgroup discovery [24, 33, 40, 46], contrast set learning [3, 4, 30], and emerging pattern mining [17, 19] identify sets of rules to describe the relationships among variables and discover interesting patterns in the data. In contrast, our work explicitly defines an objective function that scores interpretability and accuracy, and by optimizing it, we find a globally near-optimal model.…”

Section: Related Workmentioning

confidence: 99%

Interpretable Decision Sets

Lakkaraju

Bach

Leskovec

2016

Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

457

View full text Add to dashboard Cite

One of the most important obstacles to deploying predictive models is the fact that humans do not understand and trust them. Knowing which variables are important in a model’s prediction and how they are combined can be very powerful in helping people understand and trust automatic decision making systems. Here we propose interpretable decision sets, a framework for building predictive models that are highly accurate, yet also highly interpretable. Decision sets are sets of independent if-then rules. Because each rule can be applied independently, decision sets are simple, concise, and easily interpretable. We formalize decision set learning through an objective function that simultaneously optimizes accuracy and interpretability of the rules. In particular, our approach learns short, accurate, and non-overlapping rules that cover the whole feature space and pay attention to small but important classes. Moreover, we prove that our objective is a non-monotone submodular function, which we efficiently optimize to find a near-optimal set of rules. Experiments show that interpretable decision sets are as accurate at classification as state-of-the-art machine learning techniques. They are also three times smaller on average than rule-based models learned by other methods. Finally, results of a user study show that people are able to answer multiple-choice questions about the decision boundaries of interpretable decision sets and write descriptions of classes based on them faster and more accurately than with other rule-based models that were designed for interpretability. Overall, our framework provides a new approach to interpretable machine learning that balances accuracy, interpretability, and computational efficiency.

show abstract

“…This is referred to as the one versus all approach. An alternative approach, called round robin, which uses a set of 2 × 2 contingency tables representing all possible pairs of groups, has been used previously for contrast set mining [3], however subsequent research that experimented with both the round robin and one versus all approaches found that the round robin approach was not appropriate when looking for differences between two similar groups [16]. Formally, with the one versus all approach, for a contrast set X, where ∃iP (X|G i ), we determine…”

Section: Comparison Of Contrasting Groupsmentioning

confidence: 99%

GENCCS: A Correlated Group Difference Approach to Contrast Set Mining

Simeon

Hilderman

2011

Machine Learning and Data Mining in Pattern Recognition

View full text Add to dashboard Cite

Abstract. Contrast set mining has developed as a data mining task which aims at discerning differences amongst groups. These groups can be patients, organizations, molecules, and even time-lines. distinguishes one from the other. A contrast set is a conjunction of attribute-value pairs that differ significantly in their distribution across groups. The search for contrast sets can be prohibitively expensive on relatively large datasets because every combination of attribute-values must be examined, causing a potential exponential growth of the search space. In this paper, we introduce the notion of a correlated group difference (CGD) and propose a contrast set mining technique that utilizes mutual information and all confidence to select the attribute-value pairs that are most highly correlated, in order to mine CGDs. Our experiments on real datasets demonstrate the efficiency of our approach and the interestingness of the CGDs discovered.

show abstract

Contrast Set Mining for Distinguishing Between Similar Diseases

Cited by 22 publications

References 8 publications

Novel Push-Front Fibonacci Windows Model for Finding Emerging Patterns with Better Completeness and Accuracy

Novel Push-Front Fibonacci Windows Model for Finding Emerging Patterns with Better Completeness and Accuracy

Interpretable Decision Sets

GENCCS: A Correlated Group Difference Approach to Contrast Set Mining

Contact Info

Product

Resources

About