http://repository.iiitd.edu.in/xmlui/handle/123456789/626 Fri, 10 Apr 2026 20:26:20 GMT 2026-04-10T20:26:20Z http://repository.iiitd.edu.in/xmlui/handle/123456789/818 Analyzing user discussion dynamics in social media platforms Mukherjee, Arpan; Chakraborty, Tanmoy (Advisor) Understanding how users behave when they connect to social networking sites creates opportunities for better interface design, more productive studies of social interactions, and improved design of content distribution systems. Traditionally, user behavior characterization methods, based on individual features of users, are not appropriate for online networking sites. In these environments, users interact with the site and with other users through a series of multiple interfaces that let them upload and view content, choose friends, rank favorite content, subscribe to users and allow many other interactions. Different types of interactions can be observed among different types of users, and interactions on the same topic between a set of users can have different patterns as well. The motivation is to explore user behavior and the underlying conversation patterns. How a set of users react to a piece of particular news, or for a political leader how a single user changes the polarity with time or how it is constant throughout the time which can be observed in his expressive online statements (e.g., - tweets, Reddit discussion, Facebook comments, etc.). We study user behavior and its patterns in primarily three parts. (i) We present a novel quantification of conflict in an online discussion. Our measure of conflict dynamics is continuous-valued, which we validate with manually annotated ratings. Firstly, we predict the probable degree of conflict a news article will face from its audience. We employ multiple machine learning frameworks for this task using various features extracted from news articles. Secondly, given a pair of users and their interaction history, we predict if their future engagement will result in a conflict. (ii) We study how the sentiment of a user towards entities can be predicted using the tweets the user has posted so far. We consider all the available sentiments of the entities for the prediction task. (iii) In this task by using the same twitter data, we study how the sentiment of a user to an entity changes with respect to time - with time how a user changes the polarity, whether it remains same or it varies after a certain point, if so how many users had the same kind of polarity drift for an entity. Fri, 01 Jun 2018 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/818 2018-06-01T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/634 Quantum algorithms for distinguishing unitary operators Shanu; Bera, Debajyoti (Advisor) Distinguishing between unitary operators is one of the fundamental problems in the field of quantum computing. In the operator identification problem, we are given access to unknown operator U as a black-box that implements either an operator U1 or an operator U2, where U1 and U2 are arbitrary unitary operators and their operations are known to us. The goal is to determine whether U is an implementation of U1 or U2. In this thesis, two different versions of operator identification problem have been studied followed by a generalization. Firstly, we consider the case when an exact implementation of the operation of the operators U1 and U2 is given to us. We show that amplitude amplification, which is one of the important tools in quantum computing, can be used to design an efficient algorithm to solve this version of operator identification problem without error. But, in the quantum circuit theory, it may not be always possible to implement an arbitrary operator exactly and it may happen that a fabricated circuit implements a close approximation of the desired unitary operator. For the second version of the problem, we consider the case where the approximate implementation of the operation of the operators U1 and U2 is given to us; once again the goal is to design an algorithm to solve the problem. Finally, we consider a general version of the operator identification problem when the candidate set is of any size, say n. That is, U implements one of the operators present in the candidate set fU1, U2, ... , Ung and we have to identify U. We propose novel approaches to solve all these three problems in this thesis. Sun, 01 Jul 2018 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/634 2018-07-01T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/633 Study of community aware network expanion Sarika; Goyal, Vikram (Advisor); Chakraborty, Tanmoy (Advisor) Community detection has gained immense popularity in recent years. Real world networks consist of millions of nodes and edges. Groups of nodes exhibit interesting characteristics, whoseknowledge can be of great help in various fields. Moreover, networks in the real world are everevolving and it is impossible to have complete information about a network at any given time.The aim of this thesis is to study the different techniques for the exploration of the incompletenetwork available. The motive is to explore the incomplete network such that nodes whichare in the community of the known nodes are brought into the network. We build a machinelearning model to predict which node should be explored. We study four methods for selectingclustering coefficient. For identifying the communities of the nodes of the incomplete networkwe study three algorithms namely, community detection by hopcount, community detection bymaximizing modularity, and community detection by maximizing permanence. We observe thatMachine learning classifier is the best approach to maximize the recall by exploring least numberof nodes, whereas global clustering coefficient is the best approach for maintaining high precision.BFS gives better f1-score as compared to other approaches for higher budget. The communitydetection algorithms are performing equally well. Though with only a slight margin, HopCountmethod of community detection gives better results for recall and permanence maximizationgives better results for precision. Fri, 01 Jun 2018 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/633 2018-06-01T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/632 A critical study of power consumption patterns in Indian apartments Apurupa, Nagasuri Venkata; Singh, Pushpendra (Advisor); Chakravarty, Sambuddho (Advisor); Buduru, Arun Balaji (Advisor) Energy conservation plays an important role in the economic development of a country. The total share of building ’s energy consumption in the country ’s energy consumption is increasing every year in India. Thus, it is essential to study the various features that impact the energy consumption in the domestic house holds which can be used for effectivepolicy making for energy conservation and also advise individuals on how to monitor andcontrol their energy consumption. The purpose of our study is to explore the variousfactors that are affecting the total power consumption in domestic households in India.The factors include the floor level of the apartment, family size, orientation of the building,no of adults and children. The study was conducted on the faculty apartments in theIIIT-Delhi campus, India from the year 2014. The significant features were identified usingregression analysis on the data. Interaction effects between the features that impact theenergy consumption was also studied. So from our analysis, we have found out that floorlevel, orientation, family size, the presence of a working couple are having a significantimpact on the energy consumption patterns of the household. We have also developed anandroid app that will help household dwellers to monitor their usage in real time and alsoto check their historical usage and compare the trends. They can get an estimate on thenumber of units consumed and average power consumed for the selected duration of time. Sun, 01 Jul 2018 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/632 2018-07-01T00:00:00Z