Multivariate temporal data analysis ‐ a review

Author(s):  
Robert Moskovitch
2019 ◽  
Vol 13 (01) ◽  
pp. 111-133
Author(s):  
Romita Banerjee ◽  
Karima Elgarroussi ◽  
Sujing Wang ◽  
Akhil Talari ◽  
Yongli Zhang ◽  
...  

Twitter is one of the most popular social media platforms used by millions of users daily to post their opinions and emotions. Consequently, Twitter tweets have become a valuable knowledge source for emotion analysis. In this paper, we present a new framework, K2, for tweet emotion mapping and emotion change analysis. It introduces a novel, generic spatio-temporal data analysis and storytelling framework that can be used to understand the emotional evolution of a specific section of population. The input for our framework is the location and time of where and when the tweets were posted and an emotion assessment score in the range [Formula: see text], with [Formula: see text] representing a very high positive emotion and [Formula: see text] representing a very high negative emotion. Our framework first segments the input dataset into a number of batches with each batch representing a specific time interval. This time interval can be a week, a month or a day. By generalizing existing kernel density estimation techniques in the next step, we transform each batch into a continuous function that takes positive and negative values. We have used contouring algorithms to find the contiguous regions with highly positive and highly negative emotions belonging to each member of the batch. Finally, we apply a generic, change analysis framework that monitors how positive and negative emotion regions evolve over time. In particular, using this framework, unary and binary change predicate are defined and matched against the identified spatial clusters, and change relationships will then be recorded, for those spatial clusters for which a match occurs. We also propose animation techniques to facilitate spatio-temporal data storytelling based on the obtained spatio-temporal data analysis results. We demo our approach using tweets collected in the state of New York in the month of June 2014.


Author(s):  
Naonori Ueda ◽  
Futoshi Naya

Machine learning is a promising technology for analyzing diverse types of big data. The Internet of Things era will feature the collection of real-world information linked to time and space (location) from all sorts of sensors. In this paper, we discuss spatio-temporal multidimensional collective data analysis to create innovative services from such spatio-temporal data and describe the core technologies for the analysis. We describe core technologies about smart data collection and spatio-temporal data analysis and prediction as well as a novel approach for real-time, proactive navigation in crowded environments such as event spaces and urban areas. Our challenge is to develop a real-time navigation system that enables movements of entire groups to be efficiently guided without causing congestion by making near-future predictions of people flow. We show the effectiveness of our navigation approach by computer simulation using artificial people-flow data.


Author(s):  
C. Shi ◽  
J. Park ◽  
L. Manuel ◽  
M. A. Tognarelli

A well-established empirical procedure, which we refer to as Weighted Waveform Analysis (WWA), is employed to reconstruct a model riser’s response over its entire length using a limited number of strain measurements. The quality of the response reconstruction is controlled largely by identification of the participating riser response modes (waveforms); hence, mode selection is vital in WWA application. Instead of selecting a set of consecutive riser vibratory modes, we propose a procedure that automatically identifies a set of non-consecutive riser modes that can thus account for higher harmonics in the riser response (at multiplies of the Strouhal frequency). Using temporal data analysis of the discrete time-stamped samples, significant response frequencies are identified on the basis of power spectrum peaks; similarly using spatial data analysis of the sparse non-uniformly sampled data, significant wavenumbers are identified using Lomb-Scargle periodograms. Knowing the riser length, the most important wavenumber is related to a specific mode number; this dominant mode is in turn related to the dominant peak in power spectra based on the temporal data analysis. The riser’s fundamental frequency is estimated as the ratio of the empirically estimated dominant spectral frequency to the dominant mode number. Additional mode numbers are also identified as spectral peak frequencies divided by the fundamental frequency. This mode selection technique is an improvement over similar WWA procedures that rely on a priori knowledge of the risers fundamental frequency or on knowledge of physical properties and assumptions on added mass contributions. At selected target locations, we compare fatigue damage rates, estimated based on the riser response reconstructed using the WWA method with the proposed automated mode selection technique (we refer to this as “improved” WWA) and those based on the “original” WWA method (that relies on a theoretically computed fundamental natural frequency of the riser). In both cases, predicted fatigue damage rates based on the empirical methods and data at various locations (other than the target) are cross-validated against damage rates based directly on measurements at the target location. Results show that the improved WWA method, which empirically estimates the riser’s fundamental natural frequency and automatically selects significant modes of vibration, may be employed to estimate fatigue damage rates quite well from limited strain measurements.


Sign in / Sign up

Export Citation Format

Share Document