The paper describes about the development of a Named Entity Recognition (NER) system for Geological text using Conditional Random Fields (CRFs). The system makes use of the different contextual information of the words along with the variety of features that are helpful in predicting the various named entity (NE) classes. The NE tagged geological corpus was developed from the collection of scientific reports and articles on the geology of the Indian subcontinent has been used to build up the system. The training set consists of more than 2 lakh words and has been manually annotated with a NE tag set of seventeen tags. The system is able to recognize 17 classes of NEs with 75.8% Fmeasure.
Introduction of new navigation signals L2C (1227.60 MHz) and L5 (1176.45 MHz) to the existing GPS (Global Positioning System) spectrum, under the modernization program of GPS offers the improvement of position accuracy. The present study aims to understand the relative robustness of the L2C and L5 signals compared to legacy L1 C/A signal during periods of scintillations in terms of durations of cycle slips encountered from an anomaly crest location, Calcutta (22.58°N, 88.38°E geographic; magnetic dip 32°N). The data analyzed in this study were recorded during the vernal equinox of 2014 (February–April), a period of high solar activity of cycle 24. Results obtained from the comparative analyses, which are perhaps one of the first from the Indian longitude sector, indicate GPS L5 to be more robust than L1 C/A and L2C in terms of occurrence and duration of cycle slips under adverse ionospheric conditions. Furthermore, loss‐of‐lock events of duration greater than 6 s are found to be more frequent for S4 ≥ 0.6. It is found that frequency sensitivity of the GPS spectrum, in terms of occurrence of cycle slips and loss of locks are in conformity with earlier results from the equatorial region but are different from the high latitudes with respect to local time of occurrence and geomagnetic activity.
With the increasing number of mobile devices, there has been continuous research on generating optimized Language Models (LMs) for soft keyboard. In spite of advances in this domain, building a single LM for low-end feature phones as well as high-end smartphones is still a pressing need. Hence, we propose a novel technique, Optimized N-gram (Op-Ngram), an end-to-end N-gram pipeline that utilises mobile resources efficiently for faster Word Completion (WC) and Next Word Prediction (NWP). Op-Ngram applies Stupid Backoff [1] and pruning strategies to generate a light-weight model. The LM loading time on mobile is linear with respect to model size. We observed that Op-Ngram gives 37% improvement in Language Model (LM)-ROM size, 76% in LM-RAM size, 88% in loading time and 89% in average suggestion time as compared to SORTED array variant of BerkeleyLM [2]. Moreover, our method shows significant performance improvement over KenLM[3] as well.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.