A prototype-based method for classification with time constraints: a case study on automated planning

Abstract

The main goal of Nearest Prototype Classification is to reduce storage space and retrieval time of classical Instance-Based Learning (IBL) algorithms. This motivation is higher in relational data since relational distance metrics are much more expensive to compute than classical distances like Euclidean distance. In this paper, we present an algorithm to build Relational Nearest Prototype Classifiers (RNPCs). When compared with Relational Instance-Based Learning (Relational IBL or RIBL) approaches, the algorithm is able to dramatically reduce the number of instances by selecting the most relevant prototypes, maintaining similar accuracy. The number of prototypes is obtained automatically by the algorithm, although it can also be bound by the user. In this work, we also show an application of RNPC for automated planning. Specifically, we describe a modeling task where a relational policy is built following an IBL approach. This approach uses the decisions taken by a planning system as learning examples. We show that when the number of learning examples is reduced with RNPC, the resulting policy is able to scale up better than the original planning system.

A review of machine learning for automated planning

“…A solution to this problem is to post-process the stored instances to only manage the most relevant ones. Examples are REPLICA (García-Durán et al ., 2012), which extracts a set of prototypes from the instances, or DISTILL (Winner & Veloso, 2003), which builds a highly compressed instances library by generalizing and merging solution plans. Exploitation of the learned knowledge . A generalized policy can be used to directly select which action to apply for a given planning context.…”

“…CABALA (de la Rosa et al ., 2007) uses object centered solution plans, called typed sequences, for node ordering during the plan search in heuristic planning. Another instance-based learner, REPLICA, implemented a Nearest Prototype Learning (García-Durán et al ., 2012) using a relational distance inspired by the metrics used in relational data-mining. OAKPlan (Serina, 2010) uses a compact graph structure to encode planning problems.…”

Fuzzy prototype classifier based on items and its application in recommender system

“…Each prototype has a class label, and classification of each new instance is carried out by finding the closest prototype using some defined distance measures [21]. There are some variations of nearest neighbor algorithms such as nearest prototype classification (NPC) algorithms [22,23], fuzzy rough prototype selection method [24], prototype-based fuzzy clustering (PFC) algorithm [25], dissimilarity-based classifiers [26]. Yera and Martinez [27] point out that a relatively high amount of research works have proved that collaborative filtering approaches with fuzzy tools can be useful in recommendation scenarios.…”

“…NPC algorithms decrease classification time by reducing the number of instances in the data set [22]. Since the accuracy of such classification techniques is very important, Deng, et al [25] apply transfer learning to prototype-based fuzzy clustering (PFC) in order to solve the problem of limitation and scarcity of data for clustering task, and Fischer, et al [36] present simple and efficient reject options for prototype-based classification to reach a reliable decision.…”

An Approach for Prototype Generation based on Similarity Relations for Problems of Classification

“…Learning methods for problems of instance-based classification use a training set to estimate the class label(s). These learning algorithms have scalability problems when the training set size grows, so the number of training instances affects the computational cost of a method [1]; as shown in [2] the nearest neighbor rule is an example of a high computational cost method when the number of instances is big.…”