How to Upgrade Propositional Learners to First Order Logic: A Case Study

Laer, Wim Van; de, L Sophie Rodriguez

doi:10.1007/978-3-662-04599-2_10

Cited by 24 publications

(20 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most statistical relational learning algorithms come from the field of inductive logic programming (ILP) [41] and derivatives. ILP systems dealing with classification tasks typically adopt the covering approach of rule induction systems [63]. ILP suffers from the high computational complexity of the task because the algorithm has to search all the relations and all the relationships between the relations [30].…”

Section: Statistical Relational Learningmentioning

confidence: 99%

Estimating the effect of word of mouth on churn and cross-buying in the mobile phone market with Markov logic networks

Dierkes

Bichler

Krishnan

2011

Decision Support Systems

View full text Add to dashboard Cite

Much has been written about word of mouth and customer behavior. Telephone call detail records provide a novel way to understand the strength of the relationship between individuals. In this paper, we predict using call detail records the impact that the behavior of one customer has on another customer's decisions. We study this in the context of churn (a decision to leave a communication service provider) and cross-buying decisions based on an anonymized data set from a telecommunications provider.Call detail records are represented as a weighted graph and a novel statistical learning technique, Markov Logic Networks, is used in conjunction with logit models based on lagged neighborhood variables to develop the predictive model. In addition, we propose an approach to propositionalization tailored to predictive modeling with social network data. The results show that information on the churn of network neighbors has a significant positive impact on the predictive accuracy and in particular the sensitivity of churn models. The results provide evidence that word of mouth has a considerable impact on customers' churn decisions and also on the purchase decisions, leading to a 19.5% and 8.4% increase in sensitivity of predictive models.

show abstract

Section: Statistical Relational Learningmentioning

confidence: 99%

Estimating the effect of word of mouth on churn and cross-buying in the mobile phone market with Markov logic networks

Dierkes

Bichler

Krishnan

2011

Decision Support Systems

View full text Add to dashboard Cite

show abstract

“…We thus have algorithms for first order logical decision tree induction , relational distance-based clustering and prediction (Kirsten, Wrobel, and Horvath, 2001), and relational reinforcement learning (Džeroski, De Raedt, and Blockeel, 1998;Džeroski, De Raedt, and Driessens, 2001), to name a few. There is even a generic recipe for upgrading propositional learning approaches to relational settings (Van Laer and De Raedt, 2001), which involves upgrading the key notions of the propositional learner. For example, to obtain a relational distance-based approach, we need to upgrade a propositional distance measure to a relational one.…”

Section: Relational Learningmentioning

confidence: 99%

Integrating Guidance into Relational Reinforcement Learning

Driessens

Džeroski

2004

Machine Learning

View full text Add to dashboard Cite

Abstract. Reinforcement learning, and Q-learning in particular, encounter two major problems when dealing with large state spaces. First, learning the Q-function in tabular form may be infeasible because of the excessive amount of memory needed to store the table, and because the Q-function only converges after each state has been visited multiple times. Second, rewards in the state space may be so sparse that with random exploration they will only be discovered extremely slowly. The first problem is often solved by learning a generalization of the encountered examples (e.g., using a neural net or decision tree). Relational reinforcement learning (RRL) is such an approach; it makes Q-learning feasible in structural domains by incorporating a relational learner into Q-learning. The problem of sparse rewards has not been addressed for RRL. This paper presents a solution based on the use of "reasonable policies" to provide guidance. Different types of policies and different strategies to supply guidance through these policies are discussed and evaluated experimentally in several relational domains to show the merits of the approach.

show abstract

“…Relational patterns can be expressed not only in SQL, but also in first-order logic (or predicate calculus), which explains why many MRDM algorithms originate from the field of inductive logic programming (ILP) (Muggleton, 1992;De Raedt, 1992;Lǎvrac and Džeroski, 1994). Upgrading a classical data mining algorithm devised for double-entry tabular data to a relational setting is not a trivial task (Van Laer and De Raedt, 2001). For instance, it may be necessary to extend the definition of distance measure to data distributed among several tables.…”

Section: Examplementioning

confidence: 99%

A relational perspective on spatial data mining

Malerba¹

2008

IJDMMM

View full text Add to dashboard Cite

Remote sensing and mobile devices nowadays collect a huge amount of spatial data, which have to be analysed in order to discover interesting information about economic, social and scientific problems. However, the presence of a spatial dimension adds some problems to data mining tasks. The geometrical representation and relative positioning of spatial objects implicitly define spatial relationships, whose efficient computation requires a tight integration of the data mining system with the spatial DBMS. The interaction between spatially close objects causes different forms of autocorrelation, whose effect should be considered to improve the predictive accuracy of induced models and patterns. Units of analysis are typically composed of several spatial objects with different properties and their structure cannot be easily accommodated by classical double entry tabular data. In the paper, a way is shown to face these problems when a (multi-)relational data mining approach is considered for spatial data analysis. Moreover, the challenges that spatial data mining poses on current relational data mining methods are presented.

show abstract

How to Upgrade Propositional Learners to First Order Logic: A Case Study

Cited by 24 publications

References 28 publications

Estimating the effect of word of mouth on churn and cross-buying in the mobile phone market with Markov logic networks

Estimating the effect of word of mouth on churn and cross-buying in the mobile phone market with Markov logic networks

Integrating Guidance into Relational Reinforcement Learning

A relational perspective on spatial data mining

Contact Info

Product

Resources

About