V. Susheela Devi

Optimal restoration of power supply in large distribution systems in developing countries

A computer aided optimal method has been developed for the restoration of electric supply to areas isolated from the network following a fault in a distribution system. A search technique is used where the search is guided by appropriate heuristics. The optimum solution entails finding the strategy which involves the operation of minimum number of switchgear for rerouting the supply within the constraint of specified loading. This is an essential requirement in countries like India where the circuit breakers are almost always manually operated and a number of transformers and feeders operate close to their rated capacity. It pays therefore to adopt different strategies at peak load and off peak conditions since the number of breaker operations is so critical. The heuristic search that is developed is applied to a large distribution system and provides very good results. >

Combination of similarity measures for time series classification using genetic algorithms

Time series classification deals with the problem of classification of data that is multivariate in nature. This means that one or more of the attributes is in the form of a sequence. The notion of similarity or distance, used in time series data, is significant and affects the accuracy, time, and space complexity of the classification algorithm. There exist numerous similarity measures for time series data, but each of them has its own disadvantages. Instead of relying upon a single similarity measure, our aim is to find the near optimal solution to the classification problem by combining different similarity measures. In this work, we use genetic algorithms to combine the similarity measures so as to get the best performance. The weightage given to different similarity measures evolves over a number of generations so as to get the best combination. We test our approach on a number of benchmark time series datasets and present promising results.

Simultaneous feature selection and clustering using particle swarm optimization

Data clustering groups data so that data which are similar to each other are in the same group and data which are dissimilar to each other are in different groups. Since generally clustering is a subjective activity, it is possible to get different clusterings of the same data depending on the need. This paper attempts to find the best clustering of the data by first carrying out feature selection and using only the selected features, for clustering. A PSO (Particle Swarm Optimization)has been used for clustering but feature selection has also been carried out simultaneously. The performance of the above proposed algorithm is evaluated on some benchmark data sets. The experimental results shows the proposed methodology outperforms the previous approaches such as basic PSO and Kmeans for the clustering problem.

Nearest Neighbour Based Classifiers

One of the simplest decision procedures that can be used for classification is the nearest neighbour (NN) rule. It classifies a sample based on the category of its nearest neighbour. When large samples are involved, it can be shown that this rule has a probability of error which is less than twice the optimum error—hence there is less than twice the probability of error compared to any other decision rule. The nearest neighbour based classifiers use some or all the patterns available in the training set to classify a test pattern. These classifiers essentially involve finding the similarity between the test pattern and every pattern in the training set.

Parallel MCNN (pMCNN) with Application to Prototype Selection on Large and Streaming Data

Abstract The Modified Condensed Nearest Neighbour (MCNN) algorithm for prototype selection is order-independent, unlike the Condensed Nearest Neighbour (CNN) algorithm. Though MCNN gives better performance, the time requirement is much higher than for CNN. To mitigate this, we propose a distributed approach called Parallel MCNN (pMCNN) which cuts down the time drastically while maintaining good performance. We have proposed two incremental algorithms using MCNN to carry out prototype selection on large and streaming data. The results of these algorithms using MCNN and pMCNN have been compared with an existing algorithm for streaming data.

Autoregressive Model for Users’ Retweeting Profiles

Social media has become an important means of everyday communication. It is a mechanism for “sharing” and “resharing” of information. While social network platforms provide the means to users for resharing (aka retweeting), it remains unclear what motivates users to retweet. Previous studies have shown that history of users’ interaction and properties of the message are good attributes to understand the retweet behaviour of users. They however, do not consider the fact that users do not read all the blogs on their site. This results in shortcomings in the models used. We realised that simple feature engineering is also not enough to address this problem. To mitigate this, we propose an incremental model called Influence Time Content (ITC) model for predicting retweeting behavior by considering the fact that users do not read all their tweets. We have tested the effectiveness of this model by using real data from Twitter. In addition, we also investigate the parameters of the model for different classes of users. We found some interesting distinguishing patterns in retweeting behavior of users. Less active users are more topically motivated for retweeting a message than active users, who on the other hand, are social in nature.

Stochastic search techniques for post-fault restoration of electrical distribution systems

Three stochastic search techniques have been used to find the optimal sequence of operations required to restore supply in an electrical distribution system on the occurrence of a fault. The three techniques are the genetic algorithm, simulated annealing and the tabu search. The performance of these techniques has been compared. A large distribution system of over 29 substations, 2500 nodes and 120 feeders has been used.

Handwritten Digit Recognition Using Soft Computing Tools

This chapter deals with the handwritten digit recognition problem. We use a variety of classifiers for solving this problem. These classifiers include: nearest neighbour classifiers and fuzzy classifiers. A major contribution of this chapter is concerned with prototype selection for pattern classification. Genetic algorithms, simulated annealing, and tabu search are used for this purpose. The performance of various classifiers is compared based on experimental results obtained using a large data set of training and test patterns.

Overlapping Community Detection in Social Network Using Disjoint Community Detection

With increasing popularity and complexity of social networks, community detection in these networks has become an important research area. Several algorithms are available to detect overlapping community structures based on different approaches. Here we propose a two step genetic algorithm to detect overlapping communities based on node representation. First, we find disjoint communities and these disjoint communities are used to find overlapping communities. We use modularity as our optimization function. Experiments are performed on both artificial and real networks to verify efficiency and scalability of our algorithm.

Introduction to pattern recognition and machine learning

Introduction Representation Nearest Neighbour Based Classifiers Bayes Classifier Decision Trees Support Vector Machines Combination of Classifiers Clustering Application: Handwritten Digit Recognition Summary and Conclusions.