Data-driven Generation of Policies (SpringerBriefs in Computer Science)
Format: PDF / Kindle (mobi) / ePub
This Springer Brief presents a basic algorithm that provides a correct solution to finding an optimal state change attempt, as well as an enhanced algorithm that is built on top of the well-known trie data structure. It explores correctness and algorithmic complexity results for both algorithms and experiments comparing their performance on both real-world and synthetic data. Topics addressed include optimal state change attempts, state change effectiveness, different kind of effect estimators, planning under uncertainty and experimental evaluation. These topics will help researchers analyze tabular data, even if the data contains states (of the world) and events (taken by an agent) whose effects are not well understood. Event DBs are omnipresent in the social sciences and may include diverse scenarios from political events and the state of a country to education-related actions and their effects on a school system. With a wide range of applications in computer science and the social sciences, the information in this Springer Brief is valuable for professionals and researchers dealing with tabular data, artificial intelligence and data mining. The applications are also useful for advanced-level students of computer science.
0). The resulting network is the classifier function, and will, given a set of values for the action attributes, return a value in the interval Œ0; 1. We can use a classification algorithm to create a learned effect estimator. Definition 3.2. Given a classification algorithm learner, a learned effect estimator is defined to be "lrn .learner; K /.t; G/. We define the learned effect estimator to return learner.K ; G/.t/. Example 3.2. If we consider the learning algorithm to be C4.5, then "lrn .C
event KB K , if the effect estimator " is a data selection effect estimator then deciding if pEff.t; G; SCA; " .K // > 0 takes O.jK j2 / time. Proof. Consider Algorithm 2 and the analysis of its running time in the proof of Proposition 3.2. If " is not a data selection effect estimator but can be computed in polynomial time, then the loop in line 3 can be computed in O.jj E /, where E is the cost of computing " . Therefore, the total running time in this case is O.jK j E C jK j2 / 3.4
manner, where these poorly understood effects and relationships can be gleaned from the data instead of assuming that they are packaged with the input. In Chap. 1 we began by introducing this problem and the concept of event databases, where attributes are assumed to be either of type action or state; the former are those that can directly be manipulated (albeit at a certain cost and with certain probability of success), while the latter constitute those that can only be influenced indirectly.
following steps: retrieval of cases that refer to situations that are similar to the present one; reuse of such cases in the context of the present situation (this step may involve adapting the solutions); and store the new case in the knowledge base in order to apply it again in the future. For instance, a CBR approach to home repairs in the presence of a lamp that won’t turn on might 6 1 Introduction and Related Work have a set of repairs that have worked in the past: change the light bulb,
probabilistic goal describing some desired combination of actions and probability bounds with which they should be entailed from the program. The model also contains functions describing the cost of changing the environment, the effect that such changes will have on the entity being modeled, and how likely the transition is to succeed. Abductive queries in ap-programs therefore correspond to a model-heavy approach to solving the problem that we are now proposing to solve in a data-driven manner.