In this paper we present two deep-learning systems that competed at SemEval-2017 Task 4 “Sentiment Analysis in Twitter”. We participated all subtasks for English tweets, involving message-level and topic-based sentiment polarity classification quantification. use Long Short-Term Memory (LSTM) networks augmented with kinds of attention mechanisms, on top word embeddings pre-trained a big collection Twitter messages. Also, text processing tool suitable social network messages, which performs...

Rank-aware query processing has become essential for many applications that return to the user only top-k objects based on individual user's preferences. Top-k queries have been mainly studied from perspective of user, focusing primarily efficient processing. In this work, first time, we study product manufacturer. Given a potential product, which are preferences is in result set? We identify novel type, namely reverse query, manufacturers assess market and impact their products competition....

Recently, skyline queries have attracted much attention in the database research community. Space partitioning techniques, such as recursive division of data space, been used for query processing centralized, parallel and distributed settings. Unfortunately, grid-based is not suitable case a query, where allpartitions are examined at same time, since many partitions do contribute to overall set, resulting lot redundant processing.

Skyline query processing has received considerable attention in the recent past. Mainly, skyline is used to find a set of non dominated data points multidimensional dataset. While most previous work assumed centralized setting, this paper we address efficient computation subspace queries large-scale peer-to-peer (P2P) networks, where dataset horizontally distributed across peers. Relying on super-peer architecture propose threshold based algorithm, called SKYPEER, which forwards requests...

Nowadays, most applications return to the user a limited set of ranked results based on individual user's preferences, which are commonly expressed through top-k queries. From perspective manufacturer, it is imperative that her products appear in highest positions for many different otherwise product not visible potential customers. In this paper, we define novel query type, namely reverse query, covers requirement: "Given product, preferences make belong result set?." Reverse queries...

Lately the advances in centralized database management systems show a trend towards supporting rank-aware query operators, like top-k, that enable users to retrieve only most interesting data objects. A challenging problem is support queries highly distributed environments. In this paper, we present novel approach, called SPEERTO, for top-k processing large-scale peer-to-peer networks, where dataset horizontally over peers. Towards goal, explore applicability of skyline operator efficiently...

Top-k queries return to the user only k best objects based on individual preferences and comprise an essential tool for rank-aware query processing. Assuming a stored data set of preferences, reverse top-k have been introduced retrieving users that deem given database object as one their results. Reverse already attracted significant interest in research, due numerous real-life applications such market analysis product placement. Currently, most efficient algorithm computing is RTA. RTA has...

Recent technological advances have enabled both the consumption and provision of mobile services (m-services) by small, portable, handheld devices. However, devices still restricted capabilities with respect to processing, storage space, energy consumption, stable connectivity, bandwidth availability. In order address these shortcomings, a potential solution is context-awareness (by context we refer implicit information related requesting user service provider that can affect usefulness...

Top- k spatial preference queries return a ranked set of the best data objects based on scores feature in their neighborhood. Despite wide range location-based applications that rely queries, existing algorithms incur non-negligible processing cost resulting high response time. The reason is computing score object requires examining its neighborhood to find with highest score. In this paper, we propose novel technique speed up performance top-k queries. To end, mapping pairs and...

Top- k queries are widely applied for retrieving a ranked set of the most interesting objects based on individual user preferences. As an example, in online marketplaces, customers (users) typically seek products (objects) that satisfy their needs. Reversing top- leads to query type instead returns find product appealing (it belongs result preferences). In this paper, we address challenging problem processing identify m influential customers, where influence is defined as cardinality reverse...

The current approach in web searching, i.e., using centralized search engines, rises issues that question their future applicability: 1) coverage and scalability, 2) freshness, 3) information monopoly. Performing a P2P architecture consists of the actual servers has potential to tackle those issues. In order achieve desired performance as well enhancing quality relative semantic overlay networks (SONS) connecting peers storing semantically related can be employed. lack global...

In this paper, we study the generation of efficient execution plans for skyline query processing in large-scale distributed environments. such a setting, each server stores autonomously fraction data, thus all servers need to process query. An plan defines order which individual queries are processed on different servers, and influences performance processing. Querying consecutively reduces amount transferred data number queried since points obtained by one prune subsequent but also...

This paper addresses the problem of efficiently computing skyline set a relational join. Existing techniques either require to access all tuples input relations or demand specialized multi-dimensional methods generate join result. To avoid these inefficiencies, we introduce novel SFSJ algorithm that fuses identification with computation is able compute correct by accessing only subset tuples, i.e., it has property early termination. employs standard for reading and readily implementable in...

An ever-increasing number of real-life applications produce spatiotemporal data that record the position moving objects (persons, cars, vessels, aircrafts, etc.). In order to provide integrated views with other relevant sources (e.g., weather, vessel databases, etc.), this is represented in RDF and stored knowledge bases following notable features: (a) dynamic, since new spatio-temporal are recorded every second, (b) size vast can easily lead scalability issues. As a result, raises need for...

Trajectory clustering is an important operation of knowledge discovery from mobility data. Especially nowadays, the need for performing advanced analytic operations over massively produced data, such as traces, in efficient and scalable ways imperative. However, discovering clusters complete trajectories can overlook significant patterns that exist only a small portion their lifespan. In this paper, we address problem Distributed Subtrajectory Clustering highly way. The challenging because...

In typical mobile applications, users seek points of interest in their vicinity (e.g., nearby restaurants) that best match preferences. We assume a set described by combination static and dynamic attributes, mi, each associated with weighting vector wi, which expresses mi's preferences over the aforementioned attribute set. The for user correspond to results top-k query, defined is performed combined attributes. current distance between point interest. Under these assumptions, potential...

In this paper, we present a system for scalable and real-time sentiment analysis of Twitter data. The proposed relies on feature extraction from tweets, using both morphological features semantic information. For the task, adopt supervised learning approach, where train various classifiers based extracted features. Finally, design implementation architecture in Storm, which contains classification tasks, scales well with respect to input data size arrival rate. By means an experimental...