Srikumar Krishnamoorthy

Pruning strategies for mining high utility itemsets

Presents an efficient high utility mining method.Employs novel pruning strategies to limit the search space of utility mining.Compares the proposed method against a state-of-the-art utility mining method.Experimentally evaluates the system on eight real and synthetic benchmark datasets.Empirical results are found to be quite promising, especially for sparse transactional databases. High utility itemset mining problem involves the use of internal and external utilities of items (such as profits, margins) to discover interesting patterns from a given transactional database. It is an extension of the basic frequent itemset mining problem and is proven to be considerably hard and intractable. This is due to the lack of inherent structural properties of high utility itemsets that can be exploited. Several heuristic methods have been suggested in the literature to limit the large search space. This paper aims to improve the state-of-the-art and proposes a high utility mining method that employs novel pruning strategies. The utility of the proposed method is demonstrated through rigorous experimentation on several real and synthetic benchmark sparse and dense datasets. A comparative evaluation of the method against a state-of-the-art method is also presented. Our experimental results reveal that the proposed method is very effective in pruning unpromising candidates, especially for sparse transactional databases.

Linguistic features for review helpfulness prediction

Presents novel linguistic features for predicting helpfulness of online reviews.Describes a method to automatically extract linguistic features from review text.Builds a predictive model and empirically evaluates it on amazon review datasets.The model is found to be a better predictor of helpfulness for experience goods. Online reviews play a critical role in customers purchase decision making process on the web. The reviews are often ranked based on user helpfulness votes to minimize the review information overload problem. This paper examines the factors that contribute towards helpfulness of online reviews and builds a predictive model. The proposed predictive model extracts novel linguistic category features by analysing the textual content of reviews. In addition, the model makes use of review metadata, subjectivity and readability related features for helpfulness prediction. Our experimental analysis on two real-life review datasets reveals that a hybrid set of features deliver the best predictive accuracy. We also show that the proposed linguistic category features are better predictors of review helpfulness for experience goods such as books, music, and video games. The findings of this study can provide new insights to e-commerce retailers for better organization and ranking of online reviews and help customers in making better product choices.

Allocating Resources to Parallel Query Plans in Data Grids

In this paper we present a heuristic algorithm for solving the parallel query plan scheduling problem in data grids. The algorithm exploits the binary tree structure of the query plan to determine profitable allocations. It takes care of multiple forms of parallelism while allocating resources to the sub- plans. Previous approaches towards solving the problem either ignores partitioned parallelism by focusing solely on pipelined parallelism wherein the communication cost is minimized or lays more emphasis on partitioned parallelism at the cost of pipelined parallelism. The work in this paper attempts to take care of both pipelining and partitioning mechanisms while optimizing the allocation of resources to the query sub-plans. The algorithm is particularly well suited for allocation of resources to sub-plans in a bushy query plan - which is the most profitable form of query plan in distributed database query optimization.

Natural language querying over databases using cascaded CRFs

Retrieving information from relational databases using a natural language query is a challenging task. Usually, the natural language query is transformed into its approximate SQL or formal languages. However, this requires knowledge about database structures, semantic relationships, natural language constructs and also handling ambiguities due to overlapping column names and column values. We present a machine learning based natural language search system to query databases without any knowledge of Structure Query Language (SQL) for underlying database. The proposed system - Cascaded Conditional Random Field is an extension to Conditional Random Fields, an undirected graph model. Unlike traditional Conditional Random Field models, we offer efficient labelling schemes to realize enhanced quality of search results. The system uses text indexing techniques as well as database constraint relationships to identify hidden semantic relationships present in the data. The presented system is implemented and evaluated on two real-life datasets.

An integrated query optimization system for data grids

The disparate and geographically distributed data sources in an enterprise can be integrated using distributed computing technologies such as data grids. The real challenge involved in such data integration efforts is in the design and development of the distributed query processing engine that lie beneath such integrated systems. In the current literature, distributed query processing and optimization is carried out in three distinct phases namely, (1) creation of single node plan, (2) generation of parallel plan, and (3) optimal site selection for plan execution. As considering the three phases in isolation leads to sub-optimal plans, the paper proposes a new distributed query optimization model that integrates all the three phases of the query optimization. This paper also presents different heuristic approaches for solving the proposed integrated distributed query processing problem. Furthermore, the presented system is integrated with a data grid solution and several real-time experiments are conducted to demonstrate its usefulness.

Efficiently mining high utility itemsets with negative unit profits

Abstract A High Utility Itemset (HUI) mining is an important problem in the data mining literature that considers utilities of items (such as profits and margins) to discover interesting patterns from transactional databases. Several data structures, pruning strategies and algorithms have been proposed in the literature to efficiently mine high utility itemsets. Most of these works, however, do not consider itemsets with negative unit profits that provide greater flexibility to a decision maker to determine profitable itemsets. This paper aims to advance the state-of-the-art and presents a generalized high utility mining (GHUM) method that considers both positive and negative unit profits. The proposed method uses a simplified utility-list data structure for storing itemset information during the mining process. The paper also introduces a novel utility based anti-monotonic property to improve the performance of HUI mining. Furthermore, GHUM adapts key pruning strategies from the basic HUI mining literature and presents new pruning strategies to significantly improve the performance of mining. The proposed method is evaluated on a set of benchmark sparse and dense datasets and compared against a state-of-the-art method. Rigorous experimental evaluation is performed and implications of the key findings are also presented. In general, GHUM was found to deliver more than an order of magnitude improvement at a fraction of the memory over the state-of-the-art FHN method.

HMiner: Efficiently mining high utility itemsets

Abstract High utility itemset mining problem uses the notion of utilities to discover interesting and actionable patterns. Several data structures and heuristic methods have been proposed in the literature to efficiently mine high utility itemsets. This paper advances the state-of-the-art and presents HMiner, a high utility itemset mining method. HMiner utilizes a few novel ideas and presents a compact utility list and virtual hyperlink data structure for storing itemset information. It also makes use of several pruning strategies for efficiently mining high utility itemsets. The proposed ideas were evaluated on a set of benchmark sparse and dense datasets. The execution time improvements ranged from a modest thirty percent to three orders of magnitude across several benchmark datasets. The memory consumption requirements also showed up to an order of magnitude improvement over the state-of-the-art methods. In general, HMiner was found to work well in the dense regions of both sparse and dense benchmark datasets.

Sentiment analysis of financial news articles using performance indicators

Mining financial text documents and understanding the sentiments of individual investors, institutions and markets is an important and challenging problem in the literature. Current approaches to mine sentiments from financial texts largely rely on domain-specific dictionaries. However, dictionary-based methods often fail to accurately predict the polarity of financial texts. This paper aims to improve the state-of-the-art and introduces a novel sentiment analysis approach that employs the concept of financial and non-financial performance indicators. It presents an association rule mining-based hierarchical sentiment classifier model to predict the polarity of financial texts as positive, neutral or negative. The performance of the proposed model is evaluated on a benchmark financial dataset. The model is also compared against other state-of-the-art dictionary and machine learning-based approaches and the results are found to be quite promising. The novel use of performance indicators for financial sentiment analysis offers interesting and useful insights.

Efficient mining of high utility itemsets with multiple minimum utility thresholds

Abstract Mining high utility itemsets is considered to be one of the important and challenging problems in the data mining literature. The problem offers greater flexibility to a decision maker in using item utilities such as profits and margins to mine interesting and actionable patterns from databases. Most of the current works in the literature, however, apply a single minimum utility threshold value and fail to consider disparities in item characteristics. This paper proposes an efficient method (MHUI) to mine high utility itemsets with multiple minimum utility threshold values. The presented method generates high utility itemsets in a single phase without an expensive intermediate candidate generation process. It introduces the concept of suffix minimum utility and presents generalized pruning strategies for efficiently mining high utility itemsets. The performance of the algorithm is evaluated against the state-of-the-art methods (HUI-MMU-TE and HIMU-EUCP) on eight benchmark datasets. The experimental results show that the proposed method delivers two to three orders of magnitude execution time improvement over the HUI-MMU-TE method. In addition, MHUI delivers one to two orders of magnitude execution time improvement over the HIMU-EUCP method, especially on moderately long and dense benchmark datasets. The memory requirements of the proposed algorithm was also found to be significantly lower.

MAPS: Multi-Attribute Search in P2P Networks using Schema

Current peer-to-peer (P2P) infrastructures such as CAN, chord and tapestry provide scalable infrastructure for simple keyword based search. But, they cannot be applied directly in many real-life applications that use multi-attribute schema. In this paper, we present MAPS, an algorithm that supports multi-attribute schema and offers multi-attribute search capability in P2P network. The presented algorithm extends tapestry routing algorithm and is capable of supporting complete as well as partial match of multi-attribute queries. We implemented the algorithm and conducted a number of experiments to study the performance behaviour of MAPS. A comparison with other multi-attribute search methods like MAAN demonstrates the usefulness of MAPS