Behavior Of Time Series Data Research Articles

NeuroTD: A Time-Frequency Based Multimodal Learning Approach to Analyze Time Delays in Neural Activities.

Studying temporal features of neural activities is crucial for understanding the functions of neurons as well as underlying neural circuits. To this end, recent researches employ emerging techniques including calcium imaging, Neuropixels, depth electrodes, and Patch-seq to generate multimodal time-series data that depict the activities of single neurons, groups of neurons, and behaviors. However, challenges persist, including the analysis of noisy, high-sampling-rate neuronal data, and the modeling of temporal dynamics across various modalities. To address these challenges, we developed NeuroTD, a novel deep learning approach to align multimodal time-series datasets and infer cross-modality temporal relationships such as time delays or shifts. Particularly, NeuroTD integrates Siamese neural networks with frequency domain transformations and complex value optimization for inference. We applied NeuroTD to three multimodal datasets to (1) analyze electrophysiological (ephys) time series measured by depth electrodes, identifying time delays among neurons across various positions, (2) investigate neural activity and behavioral time series data derived from Neuropixels and 3D motion captures, establishing causal relationships between neural activities and corresponding behavioral activities, and (3) explore gene expression and ephys data of single neurons from Patch-seq, identifying gene expression signatures highly correlated with time shifts in ephys responses. Finally, NeuroTD is open-source at https://github.com/daifengwanglab/NeuroTD for general use.

Continuous Assessment of Function and Disability via Mobile Sensing: Real-World Data-Driven Feasibility Study.

Functional limitations are associated with poor clinical outcomes, higher mortality, and disability rates, especially in older adults. Continuous assessment of patients' functionality is important for clinical practice; however, traditional questionnaire-based assessment methods are very time-consuming and infrequently used. Mobile sensing offers a great range of sources that can assess function and disability daily. This work aims to prove the feasibility of an interpretable machine learning pipeline for predicting function and disability based on the World Health Organization Disability Assessment Schedule (WHODAS) 2.0 outcomes of clinical outpatients, using passively collected digital biomarkers. One-month-long behavioral time-series data consisting of physical and digital activity descriptor variables were summarized using statistical measures (minimum, maximum, mean, median, SD, and IQR), creating 64 features that were used for prediction. We then applied a sequential feature selection to each WHODAS 2.0 domain (cognition, mobility, self-care, getting along, life activities, and participation) in order to find the most descriptive features for each domain. Finally, we predicted the WHODAS 2.0 functional domain scores using linear regression using the best feature subsets. We reported the mean absolute errors and the mean absolute percentage errors over 4 folds as goodness-of-fit statistics to evaluate the model and allow for between-domain performance comparison. Our machine learning-based models for predicting patients' WHODAS functionality scores per domain achieved an average (across the 6 domains) mean absolute percentage error of 19.5%, varying between 14.86% (self-care domain) and 27.21% (life activities domain). We found that 5-19 features were sufficient for each domain, and the most relevant being the distance traveled, time spent at home, time spent walking, exercise time, and vehicle time. Our findings show the feasibility of using machine learning-based methods to assess functional health solely from passively sensed mobile data. The feature selection step provides a set of interpretable features for each domain, ensuring better explainability to the models' decisions-an important aspect in clinical practice.

Online Anomaly Detection for Smartphone-Based Multivariate Behavioral Time Series Data.

Smartphones can be used to collect granular behavioral data unobtrusively, over long time periods, in real-world settings. To detect aberrant behaviors in large volumes of passively collected smartphone data, we propose an online anomaly detection method using Hotelling’s T-squared test. The test statistic in our method was a weighted average, with more weight on the between-individual component when the amount of data available for the individual was limited and more weight on the within-individual component when the data were adequate. The algorithm took only an runtime in each update, and the required memory usage was fixed after a pre-specified number of updates. The performance of the proposed method, in terms of accuracy, sensitivity, and specificity, was consistently better than or equal to the offline method that it was built upon, depending on the sample size of the individual data. Future applications of our method include early detection of surgical complications during recovery and the possible prevention of the relapse of patients with serious mental illness.

An applied deep learning approach for estimating soybean relative maturity from UAV imagery to aid plant breeding decisions

For a global breeding organization, identifying the next generation of superior crops is vital for its success. Recognizing new genetic varieties requires years of in-field testing to gather data about the crop’s yield, pest resistance, heat resistance, etc. At the conclusion of the growing season, organizations need to determine which varieties will be advanced to the next growing season (or sold to farmers) and which ones will be discarded from the candidate pool. Specifically for soybeans, identifying their relative maturity is a vital piece of information used for advancement decisions. However, this trait needs to be physically observed, and there are resource limitations (time, money, etc.) that bottleneck the data collection process. To combat this, breeding organizations are moving towards advanced image capturing devices. In this paper, we develop a robust and automatic approach for estimating the relative maturity of soybeans using a time series of UAV images. An end-to-end hybrid model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) is proposed to extract features and capture the sequential behavior of time series data. The proposed deep learning model was tested on six different environments across the United States Results suggest the effectiveness of our proposed CNN-LSTM model compared to the local regression method. Furthermore, we demonstrate how this newfound information can be used to aid in plant breeding advancement decisions.

Finding Structure in Time: Visualizing and Analyzing Behavioral Time Series.

The temporal structure of behavior contains a rich source of information about its dynamic organization, origins, and development. Today, advances in sensing and data storage allow researchers to collect multiple dimensions of behavioral data at a fine temporal scale both in and out of the laboratory, leading to the curation of massive multimodal corpora of behavior. However, along with these new opportunities come new challenges. Theories are often underspecified as to the exact nature of these unfolding interactions, and psychologists have limited ready-to-use methods and training for quantifying structures and patterns in behavioral time series. In this paper, we will introduce four techniques to interpret and analyze high-density multi-modal behavior data, namely, to: (1) visualize the raw time series, (2) describe the overall distributional structure of temporal events (Burstiness calculation), (3) characterize the non-linear dynamics over multiple timescales with Chromatic and Anisotropic Cross-Recurrence Quantification Analysis (CRQA), (4) and quantify the directional relations among a set of interdependent multimodal behavioral variables with Granger Causality. Each technique is introduced in a module with conceptual background, sample data drawn from empirical studies and ready-to-use Matlab scripts. The code modules showcase each technique’s application with detailed documentation to allow more advanced users to adapt them to their own datasets. Additionally, to make our modules more accessible to beginner programmers, we provide a “Programming Basics” module that introduces common functions for working with behavioral timeseries data in Matlab. Together, the materials provide a practical introduction to a range of analyses that psychologists can use to discover temporal structure in high-density behavioral data.

SARIMA damp trend grey forecasting model for airline industry

The aim of this paper is to propose a new model that improves the Damp Trend Grey Model (DTGM) with a dynamic seasonal damping factor to forecast routes passengers demand (pax) in the air transportation industry. The model is called the SARIMA Damp Trend Grey Forecasting Model (SDTGM). In the DTGM, the damp trend factor is a static smoothing factor because it does not change over time, and therefore, it cannot capture the dynamic behavior of time series data. For this reason, the modification consists in using the trend and seasonality effects of time series data to calculate a dynamic damp trend factor as time grows. The DTGM damping factor is based on the forecasted data obtained by the GM(1,1) model; otherwise, the SDTGM calculates a seasonal damping factor based on historical data using a large amount of data points for short lead-times. The SDTGM has less uncertainty than the DTGM. The simulation results show that the SDTGM captures the seasonality effect and does not allow the forecast to exponentially grow. The SDTGM forecasts more reasonable routes pax for short lead-times when having a large amount of data points than the DTGM. The United States domestic air transport market data are used to compare the performance of the DTGM against the proposed SDTGM.

Identification of animal behavioral strategies by inverse reinforcement learning.

Animals are able to reach a desired state in an environment by controlling various behavioral patterns. Identification of the behavioral strategy used for this control is important for understanding animals’ decision-making and is fundamental to dissect information processing done by the nervous system. However, methods for quantifying such behavioral strategies have not been fully established. In this study, we developed an inverse reinforcement-learning (IRL) framework to identify an animal’s behavioral strategy from behavioral time-series data. We applied this framework to C. elegans thermotactic behavior; after cultivation at a constant temperature with or without food, fed worms prefer, while starved worms avoid the cultivation temperature on a thermal gradient. Our IRL approach revealed that the fed worms used both the absolute temperature and its temporal derivative and that their behavior involved two strategies: directed migration (DM) and isothermal migration (IM). With DM, worms efficiently reached specific temperatures, which explains their thermotactic behavior when fed. With IM, worms moved along a constant temperature, which reflects isothermal tracking, well-observed in previous studies. In contrast to fed animals, starved worms escaped the cultivation temperature using only the absolute, but not the temporal derivative of temperature. We also investigated the neural basis underlying these strategies, by applying our method to thermosensory neuron-deficient worms. Thus, our IRL-based approach is useful in identifying animal strategies from behavioral time-series data and could be applied to a wide range of behavioral studies, including decision-making, in other organisms.

Financial news predicts stock market volatility better than close price

The behaviour of time series data from financial markets is influenced by a rich mixture of quantitative information from the dynamics of the system, captured in its past behaviour, and qualitative information about the underlying fundamentals arriving via various forms of news feeds. Pattern recognition of financial data using an effective combination of these two types of information is of much interest nowadays, and is addressed in several academic disciplines as well as by practitioners. Recent literature has focused much effort on the use of news-derived information to predict the direction of movement of a stock, i.e. posed as a classification problem, or the precise value of a future asset price, i.e. posed as a regression problem. Here, we show that information extracted from news sources is better at predicting the direction of underlying asset volatility movement, or its second order statistics, rather than its direction of price movement. We show empirical results by constructing machine learning models of Latent Dirichlet Allocation to represent information from news feeds, and simple naïve Bayes classifiers to predict the direction of movements. Empirical results show that the average directional prediction accuracy for volatility, on arrival of new information, is 56%, while that of the asset close price is no better than random at 49%. We evaluate these results using a range of stocks and stock indices in the US market, using a reliable news source as input. We conclude that volatility movements are more predictable than asset price movements when using financial news as machine learning input, and hence could potentially be exploited in pricing derivatives contracts via quantifying volatility.

Modeling the Trend of Credit Card Usage Behavior for Different Age Groups Based on Singular Spectrum Analysis

Credit card holders from different age groups have different usage behaviors, so deeply investigating the credit card usage condition and properly modeling the usage trend of all customers in different age groups from time series data is meaningful for financial institutions as well as banks. Until now, related research in trend analysis of credit card usage has mostly been focused on specific group of people, such as the behavioral tendencies of the elderly or college students, or certain behaviors, such as the increasing number of cards owned and the rise in personal card debt or bankruptcy, in which the only analysis methods employed are simply enumerating or classifying raw data; thus, there is a lack of support in specific mathematical models based on usage behavioral time series data. Considering that few systematic modeling methods have been introduced, in this paper, a novel usage trend analysis method for credit card holders in different age groups based on singular spectrum analysis (SSA) has been proposed, using the time series data from the Survey of Consumer Payment Choice (SCPC). The decomposition and reconstruction process in the method is proposed. The results show that the credit card usage frequency falls down from the age of 26 to the lowest point at around the age of 58 and then begins to increase again. At last, future work is discussed.

A Time-Varying Parameter Model for Local Explosions

Locally explosive behavior is observed in many economic and financial time series when bubbles are formed. We introduce a time-varying parameter model that is capable of describing this behavior in time series data. Our proposed model can be used to predict the emergence, existence and burst of bubbles. We adopt a flexible observation driven model specification that allows for different bubble shapes and behavior. We establish stationarity, ergodicity, and bounded moments of the data generated by our model. Furthermore, we obtain the consistency and asymptotic normality of the maximum likelihood estimator. Given the parameter estimates, our filter is capable of extracting the unobserved bubble process from observed data. We study finite-sample properties of our estimator through a Monte Carlo simulation study. Finally, we show that our model compares well with noncausal models in a financial application concerning the Bitcoin/US dollar exchange rate.

Using a Complexity Approach to Study the Interpersonal Dynamics in Teacher-­Student Interactions: A Case Study of Two Teachers

In the present study, complex dynamic systems theory and interpersonal theory are combined to describe the teacher-student interactions of two teachers with different interpersonal styles. The aim was to show and explain the added value of looking at different steps in the analysis of behavioral time-series data (i.e., observations of teacher and student behaviors) that are described by Warner (1998): (1) the general level and overall coordination, (2) the presence of linear, quadratic and cubic trends in behavior, (3) the coherence and phase in cyclical trends that are superimposed on the linear, quadratic and cubic trends, and (4) the residual fluctuations, when studying the fit between teacher and student interpersonal behavior. Interactional fit is conceptualized, and described in each step of the time-series analysis, using the principle of complementarity (e.g., Kiesler, 1996). Results showed that the teacher-student interactions of the teacher with the most desirable interpersonal style largely followed the complementarity principle, whereas the interactions of the teacher with the less desirable interpersonal style did not. These results are discussed in light of the hypotheses and limitations of the study.

A novel two‐dimensional correlation coefficient for assessing associations in time series data

ABSTRACTThe widely used Pearson's correlation coefficient calculated for assessing the linear relationship between two variables might produce misleading results especially in the comparison of periodic variables. A single correlation coefficient provides a measure of the overall dependence structure and generally might not be sufficient for assessing local differences between the variables (e.g. associations between each individual year might vary in hydrologic series). The reason for this deficiency is the consideration of the averages of the whole series while ignoring the variations of the local averages (e.g. annual averages or long year averages of months) throughout the observations. This study presents a two‐dimensional horizontal (row wise) and vertical (column wise) correlation calculation approach where the compared series are considered as two‐dimensional matrices in which each row represents a sub‐period (e.g. one calendar year of the precipitation data) of the investigated time series data. The method applies a normalization procedure by considering the averages of all rows (namely local averages) for calculating the horizontal correlation and the averages of all columns for calculating the vertical correlation instead of considering the averages of the whole matrices. This enables a separate determination of the degree of relationships between the rows and columns of the compared data matrices by using the horizontal and vertical variance and covariance values that constitute the base of the two‐dimensional correlation. The method is applied on 14 different linearly varying hypothetical matrices, 6 matrices for testing the influence of seasonal and inter‐annual variations and the monthly total precipitation records of 6 stations in southwest Turkey. The results have shown that the developed correlation approach assesses the two‐dimensional behaviour of time series data like precipitation and provides a measure which enables separate assessment of the contributions from the seasonal cycle vs. inter‐annual variability in the association between two time series.

Inferring activity budgets in wild animals to estimate the consequences of disturbances

Activity budgets can provide a direct link to an animal's bioenergetic budget and is thus a valuable unit of measure when assessing human-induced nonlethal effects on wildlife conservation status. However, activity budget inference can be challenging for species that are difficult to observe and require multiple observational variables. Here, we assessed whether whalewatching boat interactions could affect the activity budgets of minke whales (Balaenoptera acutorostrata). We used a stepwise modeling approach to quantitatively record, identify, and assign activity states to continuous behavioral time series data, to estimate activity budgets. First, we used multiple behavioral variables, recorded from continuous visual observations of individual animals, to quantitatively identify and define behavioral types. Activity states were then assigned to each sampling unit, using a combination of hidden and observed states. Three activity states were identified: nonfeeding, foraging, and surface feeding (SF). From the resulting time series of activity states, transition probability matrices were estimated using first-order Markov chains. We then simulated time series of activity states, using Monte Carlo methods based on the transition probability matrices, to obtain activity budgets, accounting for heterogeneity in state duration. Whalewatching interactions reduced the time whales spend foraging and SF, potentially resulting in an overall decrease in energy intake of 42%. This modeling approach thus provides a means to link short-term behavioral changes resulting from human disturbance to potential long-term bioenergetic consequences in animals. It also provides an analytical framework applicable to other species when direct observations of activity states are not possible.

Time‐Series Characteristics and Long‐Run Equilibrium for Major Australian Office Markets

While much attention has focused on the modeling of office property markets, little emphasis has been put on distinguishing between permanent and temporary effects. This article attempts to address this issue in the context of the rental adjustment mechanism and the demand–employment relationship for major Australian central business district office markets. It is shown that, by allowing the natural vacancy rate and the work–space ratio to be endogenously determined, it offers richer model specifications that permit a partitioning between long‐run and short‐run influences. This is achieved by employing econometric techniques that examine the stochastic behavior of time series data. It is found that, while equilibrium relationships exist (between the vacancy rate and rent, and demand and employment), other macroeconomic variables are found to be relevant cyclical determinants.

Sales forecasting for thin film transistor liquid crystal display products with data clustering and an evolving neural network model

There are specific characteristics in the thin film transistor liquid crystal display (TFT-LCD) industry, such as unexpected demand fluctuation, customized products that each customer will designate a specific key component, long lead time of procurement, and short product life cycle. This research presents a genetic algorithm integrated with a neural network model for monthly sales forecasting of TFT-LCD products in Taiwan. The practical situation and its requirements are explained and two systematic approaches are discussed: (a) a K-means clustering for historic sales data and new sales is forecasted by mapping its inputs into one of these clustered data, and (b) a neural network integrated with a genetic algorithm for supervised learning of the functional behaviour of time-series data and their approximation. The evolving neural network model is applied for modelling the system's behaviour with the possibility of exploiting expert information and systematic optimization. The model has been tested and satisfying results are shown with practical data.

Comparison of periodogram tests

The periodogram ordinate is a common statistic used for the detection of sinusoidal behavior in time series data. Different statistical tests based on the periodogram ordinates have been proposed, and in this paper, Monte Carlo simulations are used to compare the power of three of these periodogram tests in their ability to detect periodic behavior under various conditions like simple or multiple cycles as well as for different frequency bases. Analyses of real world data will illustrate the strengths and weaknesses of the tests.

Generalized Precedent Logics for Resolving Insecurity Dilemmas∗

This paper proposes the development and application of a series of gradually more powerful “reflective” logical procedures to increasingly complex and realistic sets of data on collective insecurity dilemmas. Such procedures place behavioral time series data within narratively structured practical contexts. Such accounts seek grounding in the interaction between situational determinants and the reflective, historical, linguistically mediated inten‐tionality of social agents. Interrelated areas of empirical analysis will include experimentally generated behaviors and narratives from Sequential Prisioners’ Dilemmas and recently collected data on international conflict since 1945, in particular Butterworth's reports on conflict management by security‐oriented international organizations. A principal objective for developing precedental logics is to better understand, and so affect, the generation, reproduction and resolution of insecurity dilemmas.