Overlap pattern synthesis with an efficient nearest neighbor classifier

Nearest neighbor (NN) classifier is the most popular non-parametric classifier. It is a simple classifier with no design phase and shows good performance. Important factors affecting the efficiency and performance of NN classifier are (i) memory required to store the training set, (ii) classification time required to search the nearest neighbor of a given test pattern, and (iii) due to the curse of dimensionality the number of training patterns needed by it to achieve a given classification accuracy becomes prohibitively large when the dimensionality of the data is high. In this paper, we propose novel techniques to improve the performance of NN classifier and at the same time to reduce its computational burden. These techniques are broadly based on: (i) overlap based pattern synthesis which can generate a larger number of artificial patterns than the number of input patterns and thus can reduce the curse of dimensionality effect, (ii) a compact representation of the given set of training patterns called overlap pattern graph (OLP-graph) which can be incrementally built by scanning the training set only once and (iii) an efficient NN classifier called OLP-NNC which directly works with OLP-graph and does implicit overlap based pattern synthesis. A comparison based on experimental results is given between some of the relevant classifiers. The proposed schemes are suitable for applications dealing with large and high dimensional datasets like those in data mining.     

Parameterized algorithms for feedback set problems and their duals in tournaments

The parameterized feedback vertex (arc) set problem is to find whether there are k vertices (arcs) in a given graph whose removal makes the graph acyclic. The parameterized complexity of this problem in general directed graphs is a long standing open problem. We investigate the problems on tournaments, a well studied class of directed graphs. We consider both weighted and unweighted versions.     

A deterministic approach to throughput scaling in wireless networks

We address the problem of how throughput in a wireless network scales as the number of users grows. Following the model of Gupta and Kumar, we consider n identical nodes placed in a fixed area. Pairs of transmitters and receivers wish to communicate but are subject to interference from other nodes. Throughput is measured in bit-meters per second. We provide a very elementary deterministic approach that gives achievability results in terms of three key properties of the node locations. As a special case, we obtain /spl Omega/(/spl radic/n) throughput for a general class of network configurations in a fixed area. Results for random node locations in a fixed area can also be derived as special cases of the general result by verifying the growth rate of three parameters. For example, as a simple corollary of our result we obtain a stronger (almost sure) version of the /spl radic/n//spl radic/(logn) throughput for random node locations in a fixed area obtained by Gupta and Kumar. Results for some other interesting non-independent and identically distributed (i.i.d.) node distributions are also provided.     

Identification in the presence of classes of unmodeled dynamics andnoise

Identification involves obtaining a model from an a priori chosen model class(es) using finite corrupted data. The corruption may be due to several reasons ranging from noise to unmodeled dynamics, since the real system may not belong to the model class. Two popular approaches-probabilistic and set-membership identification-deal with this problem by imposing temporal constraints on the noise sample paths. We differentiate between the two sources of error by imposing different types of constraints on the corruption. If the source of corruption is noise, we model it by imposing temporal constraints on the possible realizations of noise. On the other hand, if it results from unmodeled dynamics informational constraints are imposed. Contrary to probabilistic identification where the parameters of the identified model converge to the true parameters in the presence of noise, current results in set-membership identification do not have this convergence property. Our approach leads to bridging this gap between probabilistic and set-membership identification when the source of corruption is noise. For the case when both unmodeled dynamics and noise are present, we derive consistency results for the case when the unmodeled dynamics can be described either by a linear time-invariant system or by a static nonlinearity     

Optimal sequences for CDMA under colored noise: a Schur-saddlefunction property

We consider direct sequence code division multiple access (DS-CDMA), modeling interference from users communicating with neighboring base stations by additive colored noise. We consider two types of receiver structures: first we consider the information-theoretically optimal receiver and use the sum capacity of the channel as our performance measure. Second, we consider the linear minimum mean square error (LMMSE) receiver and use the signal-to-interference ratio (SIR) of the estimate of the symbol transmitted as our performance measure. Our main result is a constructive characterization of the possible performance in both these scenarios. A central contribution of this characterization is the derivation of a qualitative feature of the optimal performance measure in both the scenarios studied. We show that the sum capacity is a saddle function: it is convex in the additive noise covariances and concave in the user received powers. In the linear receiver case, we show that the mini average power required to meet a set of target performance requirements of the users is a saddle function: it is convex in the additive noise covariances and concave in the set of performance requirements     

Vartalaap: A distributed multicast communication system

With the increasing proliferation of computer networks and distributed systems, there is a growing number of applications using multicast communication. This paper presents the Vartalaap system developed at IIT, Bombay. Vartalaap is an hierarchical distributed system for multicast communication over a network, implemented in a hardware-independent fashion. The multicast is achieved without resorting to unnecessary broadcasting of messages over the network. Issues covered in this paper include the primitives for multicast, the multicast model and the system architecture. We discuss the implementation of Vartalaap and compare it with some other systems. We conclude with a discussion on the limitations of the current implementation and directions for future work.     

The effect of different formulations on physical properties of cakes baked with microwave and near infrared-microwave combinations.

The aim of this study was to determine the effect of different formulations on color and textural characteristics of different cakes during baking in microwave and near infrared-microwave combination ovens. For comparison, cakes were also baked in conventional ovens. Color and hardness for both types of baking schemes were found to be dependent on formulation. Cakes containing Simplesse, a fat replacer consisting mostly of whey protein, baked in microwave and near infrared-microwave combination ovens were found to be the firmest cakes.