Ml Experiments with an Accuracy-eeciency Trade Oo in Unsupervised Attribute Prediction
نویسنده
چکیده
There is a lot of attention paid to parameter setting in supervised machine learning elds. Despite ML unsupervised systems also having many parameters to set, there has not as yet been much attention focused on the subject. In this report we summarise the major types of unsupervised systems and discuss the diierent concept quality measures used in unsupervised systems. An automatic parameter selecting version is built using an existing unsupervised machine learning algorithm and the wrapper method of parameter selection. This system is used to explore an accuracy-eeciency trade oo in unsupervised learning.
منابع مشابه
Exploring Eecient Attribute Prediction in Hierarchical Clustering
This work explores the feasibility of constructing hierarchical clusterings minimizing the expected cost of exploiting these clusterings for a prediction task. Particularly, we focus on gaining eeciency by means of reducing the number of features used to describe each node in the hierarchy. To explore a number of diierent hierarchical clusterings we use the Isaac clustering system, which can se...
متن کاملEecient Parallelismvs Reliable Distribution: a Trade-oo for Concurrent Computations ?
Concurrent computations should combine eeciency with reliability , where eeciency is usually associated with parallel and reliability with distributed computing. Such a desirable combination is not always possible, because of an intuitive trade-oo: eeciency requires removing redundancy from computations whereas reliability requires some redundancy. We survey a spectrum of algorithmic models (fr...
متن کاملDistributed Algorithms for Frequency Allocation: Tuning Strategies and Trade-oos
In this paper we present some of our work in progress in tuning strategies applied to two distributed algorithms that address the problem of frequency allocation for mobile networks. The tuning strategies attempt to address important aspects of mobile user access, such as connection response time, as well as the eeciency of the frequency acquisition process. In applying the strategies and exami...
متن کاملOptimization of sediment rating curve coefficients using evolutionary algorithms and unsupervised artificial neural network
Sediment rating curve (SRC) is a conventional and a common regression model in estimating suspended sediment load (SSL) of flow discharge. However, in most cases the data log-transformation in SRC models causing a bias which underestimates SSL prediction. In this study, using the daily stream flow and suspended sediment load data from Shalman hydrometric station on Shalmanroud River, Guilan Pro...
متن کاملInductive Hypothesis Validation and Bias Selection in Unsupervised Learning
This paper approaches the importance of bias selection in the context of validating Knowledge Bases (KB) obtained by inductive learning systems. We propose a framework for automatic validation of induced KBs based on the capability of shifting the bias in the inductive learning system. We claim that this framework is useful not only when the system has to validate its own results, but also when...
متن کامل