Abstract

n recent years, the importance of feature engineer-ing has been confirmed by the exceptional performance of deep learning techniques, that automate this task for some applications.For others, feature engineering requires substantial manual effort in designing and selecting features and is often tedious and non-scalable. We present AutoLearn, a regression-based feature learning algorithm. Being data-driven, it requires no domain knowledge and is hence generic. Such a representation is learnt by mining pairwise feature associations, identifying the linear or non-linear relationship between each pair, applying regression and selecting those relationships that are stable and improve the prediction performance. Our experimental evaluation on 18 UCIrvine and 7 Gene expression datasets, across different domains,provides evidence that the features learnt through our model can improve the overall prediction accuracy by 13.28%, compared to original feature space and 5.87% over other top performing models, across 8 different classifiers without using any domain knowledge.

Abstract

Computational Core for Plant Metabolomics (CCPM) is a web-based collaborative platform for researchers in the field of metabolomics to store, analyze and share their data. Metabolomics is a newly emerging field of ‘omics’ research concerned with the characterization of large numbers of metabolites using chromatography in conjunction with mass spectrometry and NMR. There is abundant volume and variety in the data, and unpredictable velocity. An interdisciplinary engagement such as this faces significant non-technical challenges solvable using a balanced approach to software management in a university setting to create an environment promoting collaborative contributions. In this paper we report on our experiences, challenges and methods in delivering a usable solution. CCPM provides a secure data repository with advanced toolsfor analysis including preprocessing, pretreatment, data filtration, statistical analysis, and pathway analysis functions; and also visualization,integration and sharing of data. As all users are not equally IT-savvy,it is essential that the user interface is robust, friendly and interactive where the user can submit and control various tasks running simultane-ously without stopping/interfering with other tasks. In each stage of its pipeline architecture, users are also allowed to upload external data that has been partially processed till the previous stage in other platforms.Use of open source softwares for development makes the maintenance and development of our modules easier than others which depend on proprietary softwares.

Abstract

n this paper, we present a generic semi-supervised Word Sense Disambiguation(WSD) method. Currently, the existingWSD methods extensively use domain re-sources and linguistic knowledge.Ourproposed method extractscontext basedlistsfrom a small sense-tagged and un-tagged training data without using do-main knowledge. Experiments on Hindiand Marathi Tourism and Health domainsshow that it gives good performance with-out using any language specific linguisticinformation except the sense IDs presentin the sense-tagged training set and workswell even with small training data by han-dling the data sparsity issue. Other ad-vantages are that domain expertise is notneeded for crafting and selecting featuresto build the WSD model and it can handlethe problem of non availability of match-ing contexts in sense-tagged training set.It also finds sense IDs of those test wordswhich are not present in sense-tag

Abstract

Word embeddings learned from text corpus can be improved by injecting knowledge from external resources, while at the same time also specializing them for similarity or relatedness. These knowledge resources (like WordNet, Paraphrase Database) may not exist for all languages. In this work we introduce a method to inject word embeddings of a language with knowledge resource of another language by leveraging bilingual embeddings. First we improve word embeddings of German, Italian, French and Spanish using resources of English and test them on variety of word similarity tasks. Then we demonstrate the utility of our method by creating improved embeddings for Urdu and Telugu languages using Hindi WordNet, beating the previously established baseline for Urdu.

Abstract

We propose an unsupervised approach to extract all competing algorithms present in a given scholarly article. The algorithm names are treated as named entities and natural language processing techniques are used to extract them.All extracted entity names are linked with their respective original papers in the reference section by our novel entity-citation linking algorithm. Then the seen citation pairs are ranked based on the number of comparison related cue-words present in the entity-citation context. We manually annotated a small subset of DBLP Computer Science conference papers and report both qualitative and quantitative results of our algorithm on it

Abstract

With the advent of statistical modeling in sports, predicting the outcome of a game has been established as a fundamental problem.Cricket is one of the most popular team games in the world. With this article, we embark on predicting the outcome of a One Day International (ODI) cricket match using a supervised learning approach from a team composition perspective. Our work suggests that the relative team strength between the competing teams forms a distinctive feature for predicting the winner. Modeling the team strength boils down to modeling individual player’s batting and bowling performances, forming the basis of our approach. We use career statistics as well as the recent performances of a player to model him. Player independent factors have also been considered in order to predict the outcome of a match. We show that the k-Nearest Neighbor (kNN) algorithm yields better results as compared to other classifiers

Abstract

In this paper we investigate the problem of user authentication based on keystroke timing pattern. We propose a simple, robust and non parameterized nearest neighbor regression based feature ranking algorithm for anomaly detection. Our approach successfully handle drawbacks like outlier detection, scale variation and prevents over fitting.Apart from using existing keystroke timing features from the dataset liked well time and flight time, other features namely big ram time and inversion ratio time are engineered as well. The efficiency and effectiveness ofour method is demonstrated through extensive comparisons with others tate-of-the-art techniques using CMU keystroke dynamics bench markage equal error rate(EER) than other proposed techniques. We achieved an average equal error rate of0.051for the user authentication task.

Abstract

We present here a data mining approach for part-of-speech (POS) tagging, an important natural language processing (NLP) task, which is a classification problem. We propose a semi-supervised associative classification method for POS tagging. Existing methods for building POS taggers require extensive domain and linguistic knowledge and resources. Our method uses a combination of a small POS tagged corpus and untagged text data as training data to build the classifier model using association rules. Our tagger works well with very little training data also. The use of semi-supervised learning provides the advantage of not requiring a large high-quality annotated corpus. These properties make it especially suitable for resource-poor languages. Our experiments on various resource-rich, resource-moderate and resource-poor languages show good performance without using any language-specific linguistic information. We note that inclusion of such features in our method may further improve the performance. Results also show that for smaller training data sizes our tagger performs better than state-of-the-art conditional random field (CRF) tagger using same features as our tagger.

Abstract

In this paper, we consider the problem of learning representations for authors from bibliographic co-authorship networks. Existing methods for deep learning on graphs, such as DeepWalk, suffer from link sparsity problem as they focus on modeling the link information only. We hypothesize that capturing both the content and link information in a unified way will help mitigate the sparsity problem. To this end, we present a novel model'Author2Vec', which learns low-dimensional author representations such that authors who write similar content and share similar network structure are closer in vector space. Such embeddings are useful in a variety of applications such as link prediction, node classification, recommendation and visualization. The author embeddings we learn are empirically shown to outperform DeepWalk by 2.35% and 0.83% for link prediction and clustering task respectively.

Abstract

Plug loads accounts for 20% to 30% of building energy consumption and has an increasing trend. Automatic plug load identification is one of the technique for effectively managing the plug load consumption.There are several studies on Non Intrusive load monitoring (NILM) but limited studies on Intrusive load monitoring (ILM). ILM is a technique that uses a low end power meter on every plug load to monitor it’s power consumption. In this paper, machine-l earning techniques are applied on low-frequency ILM data for identification of plug load and it’s state (ON/Sleep).The results show the identification accuracies are close to 98%