Abstract Automatic trading systems cope with the needs of put out emotional biases from the trading operation of public assets. These systems place orders based on a price model that forecasts the future price of an asset. Those systems, developed by edge funds and institutional investors, are not available to the public, and extensive research in this field is worth the effort. In this research, we developed a short-term price model based on a neural network and used it to forecast the near-future price direction. More in depth, we introduced the feature extraction process and parametric labeling strategy to build an ML ready dataset that includes more than 400 cryptocurrencies. The model is then validated by building a trading strategy on the two most capitalized cryptos at the time of writing: Bitcoin and Ethereum. The validation uses a trading simulation that spans six years of historical data for Bitcoin and Ethereum, including both retrospective (backtest) and prospective (forward test) evaluations. The results demonstrate that the neural network-based model exhibits a very good generalization to patterns found in historical data, enabling predictions in future data within the trading simulation. In addition, a comprehensive analysis of the importance of features was conducted to enhance the interpretability and performance of the model. Finally, we test our model in a simulated trading session; it shows that, with a simple buy-only strategy plus a stop loss, the trading system limits the draw dawn during bear markets.

This paper introduces a lightweight uncertainty estimator capable of predicting multimodal (disjoint) uncertainty bounds by integrating conformal prediction with a deep-learning regressor. We specifically discuss its application for visual odometry (VO), where environmental features such as flying domain symmetries and sensor measurements under ambiguities and occlusion can result in multimodal uncertainties. Our simulation results show that uncertainty estimates in our framework adapt sample-wise against challenging operating conditions such as pronounced noise, limited training data, and limited parametric size of the prediction model. We also develop a reasoning framework that leverages these robust uncertainty estimates and incorporates optical flow-based reasoning to improve prediction prediction accuracy. Thus, by appropriately accounting for predictive uncertainties of data-driven learning and closing their estimation loop via rule-based reasoning, our methodology consistently surpasses conventional deep learning approaches on all these challenging scenarios–pronounced noise, limited training data, and limited model size–reducing the prediction error by 2–3×.

Algorithmic trading enables the execution of orders using a set of rules determined by a computer program. Orders are submitted based on an asset’s expected price in the future, an approach well suited for high-volatility markets, such as those trading in cryptocurrencies. The goal of this study is to find a reliable and profitable model to predict the future direction of a crypto asset’s price based on publicly available historical data. We first develop a novel labeling scheme and map this problem into a Machine Learning classification problem. The model is then validated on three major cryptocurrencies through an extensive backtest over a bull, bear and flat market. Finally, the contribution of each feature to the classification output is analyzed.

In this paper, we propose an intelligent and private method to profile social network users. The scenario is constituted by the social network platform and an advertiser willing to expose its products to potentially interested users. The intelligence is in the use of text analysis services (like Wikify!) to extract knowledge from the social networks posts and get measure criteria to compare them, by using Rough Set Theory. The private layer guarantees that neither of the parties in the scenario can gain any advantage by knowing information (tastes, users topic adherence, etc...) of the other party.

The main, if not the only, income for social networks is from advertising. Social media platforms like Twitter have become a main stream communication medium to disseminate information and capture the interest of potential customers. So, it is crucial that the policy implemented to decide which ads to show in proximity of which user's posts, is the most profitable one: the ads shown should be as much as possible targeted to the user's interests. In this paper, we propose a context-aware advertising recommendation system that, analyzing the users' tweets during the timeline, interpretes the personal interests of users through orthopairs (they are equivalent to rough sets) to meet ads and users' interests at the right time.

Summary Social media is assuming a crucial role in purchasing decisions, and most companies are using social media for marketing. Making sense of the unstructured information content shared by users along the time is an emerging challenge. We advise that the dynamic nature with which trends and user's interests evolve along the timeline requires the revising of well‐assessed methods to address, for instance, recommendation provisioning, information retrieval, and so on. Time‐awareness is crucial to more effectively estimate user's interests in the future to better address social media marketing attempting to increase the traction, for instance, posting the right message at the right time. This work defines time‐aware collaborative filtering for estimating users' interest along the time in Twitter. It uses text analysis services to semantically annotate tweets' content and to track concepts considering post frequencies along the time. A model‐based approach implementing K‐Nearest Neighbors is used to estimate user's similarity representing their profile by sampling user's interest with three different techniques: Vectorial Representation, Symbolic Aggregation Approximation, and Median. We show the experimental results comparing these techniques and performing model training in different time windows. The proposed approach is used to address some social media marketing questions.

With the rapid development ot biomedical sciences, a growing amount of papers reporting new scientific findings are published and indexed in different unstructured biomedical data sources. In order to really appreciate and effectively benefit from the availability of this amount of data there is an urgent need to support the deployment of intelligent information services, such as: temporal trends and group detection, expert finding, review experts, link prediction, and so on. This need is even more stressed if we analyze dissemination activity of emerging scientific communities that are working on specific research topics in the field of biomedical science. Motivated by the fact that nanotechnologies are one of the key enabling technologies nowadays, in this paper we instantiate and contextualize the Time Aware Knowledge Extraction (TAKE) methodology, introduced in previous work, as a tool to analyze the activities of the nano-safety scientific community coordinated by the EU NanoSafety Cluster (NSC). This methodology enables us to extract timed association rules. To validate and give evidence of the goodness of these rules, a summary of the so obtained results of the analysis is provided identifying distinguishing features, and detecting emerging collaboration among the NSC's Working Groups and their members over the timeline.

Social media and networks are used by millions of people to share with their friends across the world: tastes, opinions, ideas, etc. The volume and the speed at which these data are produced make it a challenging task to discover meaningful patterns in the data. Nevertheless, very interesting business goals could be achieved collecting these data and performing analytics on social media data streams, such as: addressing marketing strategies, targeting advertisements, and so forth. We emphasize that there is a need to investigate and define suitable knowledge mining approaches to go beyond explicitly available metadata by analyzing unstructured data to provide intelligent analytics services. Specifically, in this paper we provide first results on applying OLAP analysis to multidimensional Tweet streams.

Microblog service has attracted much attention in big data analysis. Twitter statistics remark that the average number of tweets per day is greater than 100 million and many thousands of them happen every minute. It's an urgent challenge to face with such a large amount of collected tweets. This work defines an online query-focused Twitter summarization framework. It crawls and semantically indexes tweets exploiting wikification service. When a user's query is submitted, the system filters out the most relevant tweets. Then, a summarization algorithm browses the knowledge structure extracted by performing a Fuzzy Formal Concept Analysis on the given filtered tweets. Experimental results reveal good performances.

With the proliferation of different heterogeneous biomedical data sources and with the growing amount of their content available over the Web, there is, on one side, the need to support mashing and data integration and, on the other side, the more urgent need to relate literature and research results that are often enclosed in unstructured textual documents. Nowadays, ontologies have been used as a common access knowledge layer playing a crucial role to support categorized access to the information resources. Moreover, manual construction of a domain-specific ontology and content categorization is a labor intensive and a time-consuming process. This work focuses on the development of a novel biomedical ontology-driven multi-facets visualization to support categorized access to heterogeneous and unstructured biomedical data sources (e.g., PubMed, WikiGenes). Specifically, the framework relies on: knowledge extraction methodology, to automatically extract ontology exploiting the Fuzzy Formal Concept Analysis theory; and ontology matching strategy to find relation between extracted ontology and the available ones in the field of biomedicine (e.g., Ontology of Gene and Genomes, Gene Ontology, Protein Ontology). The evaluation will be shown in terms of Precision and Recall by using biomedical ontology concepts as input query to the multi-facets visualization engine.

Recently, temporal logics such as μ-calculus and Computational Tree Logic, CTL, augmented with graded modalities have received attention from the scientific community, both from a theoretical side and from an applicative perspective. In both these se

Recently, complexity issues related to the decidability of the μ-calculus, when the universal and existential quantifiers are augmented with graded modalities, have been investigated by Kupfermann, Sattler and Vardi ([19]). Graded modalities refer to the use of the universal and existential quantifiers with the added capability to express the concept of at least k or all but k, for a non-negative integer k. In this paper we study the Computational Tree Logic CTL, a branching time extension of classical modal logic, augmented with graded modalities and investigate the complexity issues with respect to the model-checking problem. We consider a system model represented by a Kripke structure K and give an algorithm to solve the model-checking problem running in time O(|K| · |φ|) which is hence tight for the problem (here |φ| is the number of temporal and boolean operators and does not include the values occurring in the graded modalities). In this framework, the graded modalities express the ability to generate a user-defined number of counterexamples to a specification φ given in CTL. However, these multiple counterexamples can partially overlap, that is they may share some behavior. We have hence investigated the case when all of them are completely disjoint. In this case we prove that the model-checking problem is both NP-hard and coNP-hard and give an algorithm for solving it running in polynomial space. We have thus studied a fragment of graded-CTL, and have proved that the model-checking problem is solvable in polynomial time.

The model checking problem for open systems has been intensively studied in the literature, for both finite-state (module checking) and infinite-state (pushdown module checking) systems, with respect to Ctl and Ctl*. In this paper, we further investigate this problem with respect to the \mu-calculus enriched with nominals and graded modalities (hybrid graded Mu-calculus), in both the finite-state and infinite-state settings. Using an automata-theoretic approach, we show that hybrid graded \mu-calculus module checking is solvable in exponential time, while hybrid graded \mu-calculus pushdown module checking is solvable in double-exponential time. These results are also tight since they match the known lower bounds for Ctl. We also investigate the module checking problem with respect to the hybrid graded \mu-calculus enriched with inverse programs (Fully enriched \mu-calculus): by showing a reduction from the domino problem, we show its undecidability. We conclude with a short overview of the model checking problem for the Fully enriched Mu-calculus and the fragments obtained by dropping at least one of the additional constructs.

We are given a line of n identical processors (finite automata) that work synchronously. Each processor can transmit just one bit of information to the adjacent processors (if any) to the left and to the right. The computation starts at time 1 with the leftmost processor in an initial state and all other processors in a quiescent state. Given the time f(n), the problem is to set (synchronize) all the processors in a particular state for the first time, at the very same instant f(n). This problem is also known as the Firing Squad Synchronization Problem and was introduced by Moore in 1964. Mazoyer has given a minimal time solution with the least number of different states (six) and very recently he has given a minimal time solution for the constrained problem in which adjacent processors can exchange only one bit. In this paper we present solutions that synchronize the line at a given time, expressed as a function of n. In particular we give solutions that synchronize at the times nlogn, n√n, n 2 and 2 n . Moreover we also show how to compose solutions in such a way to obtain synchronizing solutions for all times expressed by polynomials with nonnegative coefficients. Clearly all such solutions work also in the general case when the bit constraint is relaxed.

We study the parallel word substitution operation: in a text t, all non overlapping occurrences of a word ω are simultaneously substituted in each possible decomposition of t with respect to ω. We give necessary conditions on the reversibility of t under parallel word substitution.