Search

For predictive quantile regressions with highly persistent regressors, a conventional test statistic suffers from a serious size distortion and its limiting distribution relies on the unknown persistence degree of predictors. This paper proposes a double-weighted approach to offer a robust inferential theory across all types of persistent regressors. We first estimate a quantile regression with an auxiliary regressor, which is generated as a weighted combination of an exogenous random walk process and a bounded transformation of the original regressor. With a similar spirit of rotation in factor analysis, one can then construct a weighted estimator using the estimated coefficients of the original predictor and the auxiliary regressor. Under some mild conditions, it shows that the self-normalized test statistic based on the weighted estimator converges to a standard normal distribution. Our new approach enjoys a nice property that it can reach the local power under the optimal rate T with nonstationary predictor and squared root of T for stationary predictor, respectively. More importantly, our approach can be easily used to characterize mixed persistence degrees in multiple regressions. Simulations and empirical studies are provided to demonstrate the effectiveness of the newly proposed approach. The heterogenous predictability of US stock returns at different quantile levels is reexamined.

Facilitated identification of predictable words during online reading has been attributed to the generation of predictions about upcoming words. But highly predictable words are relatively infrequent in natural texts, raising questions about the utility and ubiquity of anticipatory prediction strategies. This study investigated the contribution of task demands and aging to predictability effects for short natural texts from the Provo corpus. The eye movements of 49 undergraduate students (mean age 21.2) and 46 healthy older adults (mean age 70.8) were recorded while they read these passages in two conditions: (i) ‘reading for meaning’ to answer occasional comprehension questions; (ii) ‘proofreading’ to detect ‘transposed letter’ lexical errors (e.g., clam instead of calm) in intermixed filler passages. The results suggested that the young adults, but not the older adults, engaged anticipatory prediction strategies to detect semantic errors in the proofreading condition, but neither age group showed any evidence of costs of prediction failures. Rather, both groups showed facilitated reading times for unexpected words that appeared in a high constraint within-sentence position. These findings suggest that predictability effects for natural texts reflect partial, probabilistic expectancies rather than anticipatory prediction of specific words.

Option-based measures can predict underlying stock returns, due to differences in price discovery and price pressure effects between options and underlying stocks. We investigate stock return predictability by various option price-based measures using REITs. REITs are more transparent and efficiently priced than general stocks, but REIT options are less liquid. Consistent with the model of Easley, O’Hara, and Srinivas (1998), most of the option price-based measures do not significantly forecast REIT stock returns, but changes in option implied volatilities are robust and significant return predictors. We provide further evidence supporting the informed trading channel instead of the price pressure effects.

We study the equilibrium implications of a multi-asset economy in which asset managers are subject to different benchmarks, and demonstrate how heterogeneous benchmarking generates a mechanism through which fundamental shocks propagate across assets. Fluctuations in asset managers' capital invested for benchmarking purposes, scaled by the size of the economy, induce price pressure that can result in negative spillovers across asset returns. We highlight the economic significance of these benchmarking-induced spillovers by analyzing shock elasticities and cross-elasticities of price-dividend ratios, and characterize a rich structure of asset price comovements within and across benchmarks. Heterogeneous benchmarking also induces return predictability, generating both reversal and momentum.

This article investigates the predictability of dry bulk shipments' physical shipping costs while testing the predictive significance of derivative products. Accordingly, a comprehensive grid search procedure is needed to simulate combinations of model structures subject to a cross-validation process. In this regard, the intelligent model search engine (IMSE) is implemented as the search algorithm. The IMSE algorithm is selected due to its broad model space naturally consisting most of the traditional econometric model structures and advanced features, such as sample size optimization (also detects structural breaks), lag structure optimization (also reflects seasonality), among others. In contrast to the previous studies in the shipping market research, IMSE executes a relaxation of statistical significance criteria to target the out of sample predictive accuracy (instead of the goodness of fit). Sparsification is retained by utilizing a dropout procedure in line with the cross-validation process. The empirical results indicate the fact that predictive accuracy can be achieved or improved with a shorter sample period. More than half of the entire dataset has not been utilized in the final selections for most simulations. One-month and two-month maturity derivative contracts are tested for predictive features, while none of them is selected for the sparse model selections.

This paper investigates spillovers from foreign economies to the U.S. through changes in longterm Treasury yields. We document a decline in the contribution of U.S. domestic news to the variance of long-term Treasury yields and an increased importance of overnight yield changes—a rough proxy for the contribution of foreign shocks to U.S. yields—over the past decades. Using a model that identifies U.S., Euro area, and U.K. shocks that move global yields, we estimate that foreign (non-U.S.) shocks account for at least 20 percent of the daily variation in long-term U.S. yields in recent years. We argue that spillovers occur in large part through bond term premia by showing that a low level of foreign yields relative to U.S. yields predicts a decline in distant forward U.S. yields and higher returns on a strategy that is long on a long-term Treasury security and short on a long-term foreign bond.

Participant selection is a fundamental research issue in Mobile Crowdsensing (MCS). Previous approaches commonly assume that adequately long periods of candidate participants' historical mobility trajectories are available to model their patterns before the selection process, which is not realistic for some new MCS applications or platforms. The sparsity or even absence of mobility traces will incur inaccurate location prediction, thus undermining the deployment of new MCS applications. To this end, this paper investigates a novel problem called From-Scratch MCS (FS-MCS for short), in which we study how to intelligently select participants to minimize such cold-start effect. Specifically, we propose a novel framework based on reinforcement learning, named RL-Recruiter+. With the gradual accumulation of mobility trajectories over time, RL-Recruiter+ is able to make a good sequence of participant selection decisions for each sensing slot. Compared to its previous version, RL-Recruiter, Re-Recruiter+ jointly considers both the previous coverage and current mobility predictability when training the participant selection decision model. We evaluate our approach experimentally based on two real-world mobility datasets. The results demonstrate that RL-Recruiter+ outperforms the baseline approaches, including RL-Recruiter under various settings.

Intuitively, one would expect a more skillful forecast if predicting weather averaged over one week instead of the weather averaged over one day, and similarly for different spatial averaging areas. However, there are few systematic studies of averaging and forecast skill with modern forecasts, and it is therefore not clear how much improvement in forecast performance is produced via averaging. Here we present a direct investigation of averaging effects, based on data from operational numerical weather forecasts. Data is analyzed for precipitation and surface temperature, for lead times of roughly 1 to 7 days, and for time- and space-averaging diameters of 1 to 7 days and 100 to 4500 km, respectively. For different geographic locations, the effects of time- or space-averaging can be different, and while no clear geographical pattern is seen for precipitation, a clear spatial pattern is seen for temperature. For temperature, in general, time averaging is most effective near coastlines, also effective over land, and least effective over oceans. Based on all locations globally, time averaging was less effective than one might expect. To help understand why time averaging may sometimes be minimally effective, a stochastic model is analyzed as a synthetic weather time series, and analytical formulas are presented for the decorrelation time. In effect, while time averaging creates a time series that is visually smoother, it does not necessarily cause a substantial increase in the predictability of the time series.

Recent research has demonstrated the adaptiveness of variability in emotion regulation (ER) by showing that variability between and, when controlled for depression, within ER strategies as assessed via the standard deviation was associated with less negative affect. We first replicated associations with negative affect by using the relative standard deviation, which is less confounded with the mean. Second, following research on affect dynamics, we extended this line of research by examining five additional ER dynamic measures covering ER instability, inertia, predictability, differentiation, and diversity. Reanalyzing data from five ambulatory assessment data sets (N = 717), we found that (a) the eight ER dynamic measures loaded on five factors that explained unique variance, (b) most ER dynamic measures had good reliabilities, and (c) between-strategy mean endorsement was positively, whereas between-strategy variability and ER predictability were negatively associated with negative affect. These results suggest that the variable but predictable use of emotion regulation strategies in daily life is beneficial for individuals' affective well-being in daily life. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Predictable hardware cache coherence is a viable shared data communication mechanism between cores for multicore real-time platforms. Prior works have established that predictable hardware cache coherence protocols offer significant performance advantages over alternative predictable data communication mechanisms while ensuring predictability. Unlike alternative predictable data communication mechanisms, designing predictable cache coherence protocols is non-trivial as it requires detailed understanding of the impact of different memory activity patterns to shared data for predictable and coherent data communication. Furthermore, designing predictable cache coherence protocols that deliver high average-case performance is even more challenging as it entails identifying opportunities such that a core’s access to a data is not stalled in the presence of interleaving memory activity from other cores to the same data. To this end, we present SYNTHIA, an open and automated tool for synthesizing predictable and high-performance snooping bus-based cache coherence protocols for multi-core platforms deployed in realtime systems. SYNTHIA automates the complex analysis associated with designing predictable and high-performance cache coherence protocols, and constructs the complete protocol implementation (coherence states and transitions) that achieve predictability and performance. We use SYNTHIA to construct complete protocol implementations from simple specifications of common protocols (MSI, MESI, and MOESI protocols) and a predictable variant of the MESIF cache coherence protocol, which was recently found to be deployed in an existing multi-core platform designed for real-time platforms. We validated the correctness, predictability, and performance guarantees of the generated protocol implementations from SYNTHIA using manually implemented versions, and a micro-architectural simulator.

International audience; We explore the characteristics of 96190 wet events (WE) defined as consecutive 3-hourly rainfall >= 1 mm/3h from a network of 34 stations across Bangladesh. Nearly 60% (5%) of wet events last = 15) hours. The WEs are dynamically clustered into four "canonical" storm types (ST), mostly discretized by their duration, but also their mean and maximal intensity. While durations, total amounts and wet contiguous areas of WEs are positively related, their mean intensity is nearly independent of them. ~ 60% of WEs are associated with ST#1, that is short and small WEs and very low rainfall amounts (usually < 10 mm), ~15% of WEs are associated with either (ST#2) short/small WEs but with intense rainfall, probably mostly related to scattered thunderstorms, or (ST#3) longer/larger WEs but with less intense rainfall. The last ST (ST#4) is rare (~ 6%), related to very long durations and large wet areas, and includes the wettest WEs. It is especially frequent over southeastern Bangladesh. ST#2 to #4 contribute almost equally to the local-scale total amount of rainfall (27-29% each in mean) while ST#1, despite its individual low rainfall amount, still includes ~ 15% of it. ST#2 (ST#4) is related to the highest probability of occurrence of 3-hourly (daily) extremes. ST#4 occurrence is the most impacted by synoptic Indian lows/depressions as well as the main modes of intraseasonal variation, while ST#1 and #2 are also significantly impacted by intraseasonal modes but in reverse manner than ST#4.

This paper performs an out-of-sample comparison of linear factor asset pricing models from an economic perspective under predictability. I assess the economic value added of several factor models when a Bayesian investor is faced with a portfolio allocation problem whereby each model imposes cross-sectional restrictions on the parameters of a predictive stock return regression. The empirical framework explicitly accounts for investor skepticism about the model, i.e., mispricing uncertainty. Using several US portfolios as test assets, I find that the q5 model of Hou et al. (2020), as well as the behavioral factor models of Stambaugh and Yuan (2017) and Daniel et al. (2020) outperform competing models across investment horizons. At the longest evaluated horizon (one year), a benchmark portfolio built using historical data produces larger portfolio gains than all the factor models, but in the short run (at the one-month horizon), their performance is comparable.

Deep neural networks (DNNs) have been widely used in many Cyber-Physical Systems (CPS). However, it is still challenging work to deploy DNNs in real-time systems. In particular, the execution time of DNN inference must be predictable, s.t. it could be known whether the run-time inference can complete within a required timing constraint. Moreover, the timing constraints may change dynamically with the run-time environment in many embedded applications, such as autonomous cars. A possible way to meet such dynamic real-time requirements is to execute different sub-networks of a DNN at run-time. However, improper construction of sub-networks may not only introduce unpredictable inference time, s.t. the real-timing constraints could be violated unexpectedly, but also has poor compatibility with the well-optimized machine learning framework (e.g., TensorFlow). In this paper, we study the predictability when executing different sub-networks of a DNN. In particular, we present a feature-wise run-time adaptation framework for DNN inference, which is implemented and validated on NVIDIA Jetson TX2 and Nano with TensorFlow. The experimental results show that our method can achieve predictable inference time in comparison with the state-of-the-art methods.

The paper examines statistical and economic evidence of out-of-sample bond return predictability for a real-time Bayesian investor who learns about parameters, hidden states, and predictive models over time. We find some statistical evidence using information contained in forward rates. However, such statistical predictability can hardly generate any economic value for investors. Furthermore, we find that strong statistical and economic evidence of bond return predictability from fully-revised macroeconomic data vanishes when real-time macroeconomic information is used. We also show that highly levered investments in bonds can improve short-run bond return predictability.

Corporate bond prices are slow to respond to default risk and interest rate shocks, as proxied by firm-level stock returns and Treasury returns, respectively. Furthermore, the underreaction is risk-specific: bonds with better credit quality underreact more to default risk, while those with worse quality underreact more to interest rates. The underreactions imply substantial out-of-sample return predictability, and investors appear to be leaving too much money on the table. The results are consistent with behavioral inattention models in which investors endogenously allocate more attention to payoff-relevant (or salient) risks, and they are not explained by traditional trading friction mechanisms.

We add to the emerging literature on empirical asset pricing in the Chinese stock market by building and analyzing a comprehensive set of factors with 1,160 signals for return prediction. Using various machine learning algorithms, we investigate which signals dominate in the Chinese market, a market characterized by a large proportion of retail investors with speculative motives, state-owned firms, and short-sales restrictions. Contrary to studies for the U.S. market, liquidity and fundamental factors emerge as the most important predictors, while price trend signals are less significant. We find that retail investors' dominating presence positively affects short-term predictability, particularly for small stocks. Another feature that distinguishes the Chinese from the U.S. market is the high predictability of large stocks and state-owned enterprises over longer horizons. Our portfolio analysis shows that this overall increased predictability leads to significantly higher out-of-sample performance than in other markets, which remains economically significant after transaction costs.

Prior research documents that asset growth is negatively associated with future firm performance. In contrast, we show that growth financed by product market stakeholders (i.e., “operating growth”) is positively associated with future firm performance. Investors and security analysts underestimate the positive effects of operating growth on future performance, resulting in return predictability and overly pessimistic earnings forecasts for firms with high operating growth. Future stock returns largely concentrate around subsequent earnings announcements with declining magnitudes, consistent with the error-in-expectation explanation. Results from cross-sectional tests further support the hypothesis that operating growth signals high future performance but investors underreact to it.

We exploit information in sovereign CDS spreads and the international trade network to provide causal evidence of the propagation of global economic shocks. We show that trade links are an important source of shock transmission using the natural experiments of the Japanese tsunami and the COVID-19 lockdown in China. We then confirm more general and gradual information flows along the trade network by showing extensive country-level credit/equity cross-sectional return predictability. News about country fundamentals flows primarily from importers to exporters, depends on both direct and indirect links in the trade network, and is magnified by the exporting country's financial vulnerability.

Using a broad selection of 53 carbon (EUA) related, commodity and financial predictors, we provide a comprehensive assessment of the out-of-sample (OOS) predictability of weekly European carbon futures return. We assess forecast performance using both statistical and economic value metrics over an OOS period spanning from January 2013 to May 2018. Two main types of dimension reduction techniques are employed: (i) shrinkage of coefficient estimates and (ii) factor models. We find that: (1) these dimension reduction techniques can beat the benchmark significantly with positive gains in forecast accuracy, despite very few individual predictors being able to; (2) forecast accuracy is sensitive to the sample period, and only Group-average models and Commodity-predictors tend to beat the benchmark consistently; the Group-average models can improve both the prediction accuracy and stability significantly by averaging the predictions of All-predictors model and the benchmark. Further, to demonstrate the usefulness of forecasts to the end-user, we estimate the certainty equivalent gains (economic value) generated. Almost all dimension reduction techniques do well especially those which apply shrinkage alone. We find including All-predictors and Group-average variable sets achieve the highest economic gains and portfolio performance. Our main results are robust to alternative specifications.

This paper analyses straightforward Bonferroni adjustments to critical values of in-sample tests of predictability, when data mining is used to search across models. Unlike conventional pseudo-out-of-sample tests, these in-sample tests have stable family-wise error rates (FWERs) when searching for models that predict well. Furthermore, when data mining, these in-sample tests have more power than pseudo-out-of-sample tests for identifying true predictability.

The SubNiño4 index based on the subsurface potential temperature around the thermocline beneath the west Pacific warm pool, the Niño 4 region, is examined as a long-range indicator of the surface El Niño–Southern Oscillation (ENSO) and ENSO-driven atmospheric response. The SubNiño4 index captures the evolution of subsurface ocean heat content between the El Niño and La Niña phases of the ENSO cycle, allowing it to serve as a long-range indicator of surface ENSO and hence also many ENSO-driven atmospheric anomalies. The SubNiño4 index has more temporally stable correlations with Niño 3.4 than the widely used western equatorial Pacific warm-water volume indicator. For a lead time of the order of 12 months, Niño 3.4 correlations afforded by the lead observed SubNiño4 index become similar to and can exceed those produced by typical dynamical ENSO predictions. The value and viability of the SubNiño4 index as a simple statistical long-range indicator of ENSO-driven atmospheric response is shown for regional precipitation anomalies throughout the Tropics and fires in Continental and Maritime Southeast Asia.

We extend existing studies by considering the higher-order moments relationships among crude oil, gold, and Bitcoin markets. Using high-frequency data from December 2, 2014 to June 10, 2018, we analyze spillovers in jumps and realized second, third, and fourth moments among crude oil, gold, and Bitcoin markets via Granger causality and generalized impulse response analyses. Results suggest evidence of predictability and emphasize, among others, the need of jointly modeling linkages across those three markets with higher-order moments; otherwise, inaccurate risk assessment and investment inferences may arise. The responses of realized volatility shocks are generally positive. Further analyses indicate evidence of a weaker relationship between gold and crude oil and Bitcoin and crude oil compared to the relationship between Bitcoin and gold. Practical implications are also discussed.

We revisit the estimation algorithm of Pettenuzzo and Timmermann (2011) and show how to apply the posterior simulation test of Geweke (2004) to locate and correct an error in the original posterior sampling algorithm. The main modification for the new algorithm is the introduction of a Metropolis–Hasting step to draw the precision parameters of the return and predictor equations, taking into account the correlation between the error terms of the two equations. We find that the modification of the original algorithm has very minor effects on the empirical results reported in Pettenuzzo and Timmermann (2011). In particular, for both the dividend yield and Treasury bill predictors the updated algorithm finds that the models best supported by the data, as measured by the SIC, match those reported in the original paper.

Bias correction in predictive regressions is known to reduce the empirical size problems of OLS-based predictability tests with persistent predictors. This paper shows that bias correction is also achieved in the context of the extended instrumental variable (IVX) predictability testing framework introduced by Kostakis et al. (2015). To be specific, new IVX-based statistics subject to a bias correction analogous to that proposed by Amihud and Hurvich (2004) are introduced. Four important contributions are provided: first, we characterize the effects that bias-reduction adjustments have on the asymptotic distributions of the IVX test statistics in a general context allowing for short-run dynamics and heterogeneity; second, we discuss the validity of the procedure when predictors are stationary as well as near-integrated; third, we conduct an exhaustive Monte Carlo analysis to investigate the small in- and out-of-sample properties of the test procedures and their sensitivity to distinctive features that characterize predictive regressions in practice, such as strong persistence, endogeneity, and non-Gaussian innovations; and fourth, we provide an analysis of real estate return and rent growth predictability in 19 OECD countries.

We propose target date funds modified to exploit stock return predictability driven by the variance risk premium. The portfolio rule of these tactical target date funds (TTDFs) is extremely simplified relative to the optimal one, making it easy to implement and to communicate to investors. We show that saving for retirement in TTDFs generates economically large welfare gains, even after we introduce turnover restrictions and transaction costs, and after taking into account parameter uncertainty. This predictability also appears to be uncorrelated with individual household risk, suggesting that households are in a prime position to exploit it. This paper was accepted by Tomasz Piskorski, finance.

This article reviews Fuqing Zhang’s contributions to mesoscale atmospheric science from research to mentoring to academic service over his 20-year career. His fundamental scientific contributions are highlighted on predictability, data assimilation, and dynamics of high impact weather, especially gravity wave and tropical cyclones. His extremely generous efforts in efficiently transmitting to the community the new scientific knowledge and ideas, through mentoring, interacting, workshop organizing, and reviewing are summarized. Special appreciations were given to his tremendous contributions to the development of mesoscale meteorology in China and the education of Chinese graduate students and young scientists.

Using historical data on postwar financial crises around the world, we show that the combination of rapid credit and asset price growth over the prior three years, whether in the nonfinancial business or the household sector, is associated with a 40% probability of entering a financial crisis within the next three years. This compares with a roughly 7% probability in normal times, when neither credit nor asset price growth is elevated. Our evidence challenges the view that financial crises are unpredictable “bolts from the sky” and supports the Kindleberger-Minsky view that crises are the byproduct of predictable, boom-bust credit cycles. This predictability favors policies that lean against incipient credit-market booms.

Different from the existing train delay studies that had strived to explore sophisticated algorithms, this paper focuses on finding the bound of improvements on predicting multi-scenario train delays with different machine learning methods. Motivated by the observation of deep learning methods failing to improve the prediction performance if the delay occurs rarely, we present a novel augmented machine learning approach to improve the overall prediction accuracy further. Our solution proposes a rule-driven automation (RDA) method, including a delay status labeling (DSL) algorithm, and the resilience of section (RSE) and resilience of station (RST) indicators to generate the forecast for train delays. The experiment results demonstrate that the Random Forest based implementation of our RDA method (RF-RDA) can significantly improve the generalization ability of multivariate multi-step forecast models for multi-scenario train delay prediction. The proposed solution surpasses state-of-art baselines based on real-world traffic datasets, which treat various real-time delays differently. Even when the predictability of conventional deep learning methods decreases, the performance of our method is still acceptable for practical use to provide accurate forecasts.

Coordinated operation of renewable resources and storage devices can reduce the undesirable effects of poor predictability of renewable producers. This article proposes a three-stage hybrid robust/stochastic framework to model the coordinated operation of wind producers and compressed air energy storage in the form of a mixed-integer linear programming problem. The proposed algorithm derives robust offering curves to participate in the day-ahead market. In addition to the day-ahead market, intraday and balancing markets are considered for reducing the negative effects of poor predictability of wind speed. In the proposed hybrid framework, intraday and balancing market prices as well as wind speed are modeled by proper scenarios, while robust optimization is used to model the day-ahead price uncertainty. Besides, the risks of stochastic parameters are modeled by the conditional value at risk (CVaR). Based on the results of robust optimization, the uncertainty of the day-ahead market price reduces the aggregator profit by 21.3% in the worst case, whereas CVaR results show a 13.2% profit reduction due to the worst-case scenario realization in stochastic parameters.

The influence of El Niño Southern Oscillation (ENSO) on Colombia's hydrological variables has been shown in different studies. Most of the methodologies implemented have identified linear relationships and have associated the warm (cold) phase called El Niño (La Niña) with negative (positive) rainfall and streamflow anomalies. One of the most adverse impacts founded is the reduction in water supply during the warm phase. Therefore, it is necessary to study the linkage between ENSO and precipitation variability for efficient management of water resources. Consequently, the present paper has two purposes. The first one is to explore nonlinear correlations of the ENSO-precipitation relationship, particularly for specific regions where the freshwater resources have been significantly reduced during El Niño events. The second one is to identify which indices will enable in improving the predictability of hydro-climatological variables. The research was based on the wavelet coherence analysis of monthly precipitation time series from 1981-2016 and the ENSO indices for the same period. The results show that ENSO events influence the precipitation as periods of rainfall deficit or excess. Also, precipitation is organized in bands and that the 2–8-year scales explain most of their variance. The most significant sectors are those that cover El Niño events. In contrast, sectors are smaller when La Niña episodes. Then impacts on precipitation tend to be greater for warm events. Results also allowed to identify that El Niño 3, Niño 3,4, ONI, and BEST indices can be good indicators for forecasting work in these specific places. The use of two kinds of data, one in situ and the other from CHIRPS program, allows to establish the feasibility of using data from satellite origin in regions without enough information; the results showed that CHIRPS data tend to report fewer anomalies than data in situ. However, the coherence structure is similar, but in periods between 36 and 48 months, there were discrepancies of pi/4 in the phase difference, that is, between 3 and 6 months of difference in lags calculated with each database.

Distributed denial of service (DDoS) attacks remain challenging to mitigate in existing systems, including in-home networks that comprise different Internet of Things (IoT) devices. In this paper, we present a DDoS traffic detection model that uses a boosting method of logistic model trees for different IoT device classes. Specifically, a different version of the model will be generated and applied for each device class, since the characteristics of the network traffic from each device class may have subtle variation(s). As a case study, we explain how devices in a typical smart home environment can be categorized into four different classes (and in our context, Class 1 -very high level of traffic predictability, Class 2 -high level of traffic predictability, Class 3 -medium level of traffic predictability, and Class 4 -low level of traffic predictability). Findings from our evaluations show that the accuracy of our proposed approach is between 99.92% and 99.99% for these four device classes. In other words, we demonstrate that we can use device classes to help us more effectively detect DDoS traffic.

This paper investigates the spatiotemporal characteristics and predictability of the emerging modern traffic behavior, ridesourcing. We collect a comprehensive data set of Didi ridesourcing cars on a large geographical scale of a capital city in China, including both the temporal order information and the GPS-recorded spatial trajectories. To extract the features of this kind of traffic behavior, we construct a large-scale network by considering every traffic flow of the orders. Therein, a driver consecutively visiting different regions of the city connects the relationship of these sites. The weighted ridesourcing network shows a consistency of the distribution of trip orders and the Clark model for population distribution. The network also has spatial and temporal features with power laws, sometimes with exponential truncations and log-normal distributions. Furthermore, we propose a general analytical method to quantify the predictability of this kind of behavior by calculating the entropy at a collective level, which can be extended to quantify other traffic behaviors. Finally, by considering the traffic congestion factor, we propose a better neural network based model for predicting dwelling time of the ridesourcing behavior. We suggest that the traffic behavior of ridesourcing cars indicates specific non-Markovian characteristics, which can be systematically analyzed from the viewpoint of network sciences.

Visual search is a fundamental element of human behavior and is predominantly studied in a laboratory setting using static displays. However, real-life search is often an extended process taking place in dynamic environments. We have designed a dynamic-search task in order to incorporate the temporal dimension into visual search. Using this task, we tested how participants learn and utilize spatiotemporal regularities embedded within the environment to guide performance. Participants searched for eight instances of a target that faded in and out of a display containing similarly transient distractors. In each trial, four of the eight targets appeared in a temporally predictable fashion with one target appearing in each of four spatially separated quadrants. The other four targets were spatially and temporally unpredictable. Participants' performance was significantly better for spatiotemporally predictable compared to unpredictable targets (Experiments 1-4). The effects were reliable over different patterns of spatiotemporal predictability (Experiment 2) and primarily reflected long-term learning over trials (Experiments 3, 4), although single-trial priming effects also contributed (Experiment 4). Eye-movement recordings (Experiment 1) revealed that spatiotemporal regularities guide attention proactively and dynamically. Taken together, our results show that regularities across both space and time can guide visual search and this guidance can primarily be attributed to robust long-term representations of these regularities. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Frequency-dependent (FD) selection, whereby fitness and selection depend on the genetic or phenotypic composition of the population, arises in numerous ecological contexts (competition, mate choice, crypsis, mimicry, etc.) and can strongly impact evolutionary dynamics. In particular, negative frequency-dependent selection (NFDS) is well known for its ability to potentially maintain stable polymorphisms, but it has also been invoked as a source of persistent, predictable frequency fluctuations. However, the conditions under which such fluctuations persist are not entirely clear. In particular, previous work rarely considered that FD is unlikely to be the sole driver of evolutionary dynamics when it occurs, because most environments are not static but instead change dynamically over time. Here, we investigate how FD interacts with a temporally fluctuating environment to shape the dynamics of population genetic change. We show that a simple metric introduced by Lewontin, the slope of frequency change against frequency near equilibrium, works as a key criterion for distinguishing microevolutionary outcomes, even in a changing environment. When this slope D is between 0 and –2 (consistent with the empirical examples we review), substantial fluctuations would not persist on their own in a large population occupying a constant environment, but they can still be maintained indefinitely as quasi-cycles fueled by environmental noise or genetic drift. However, such moderate NFDS buffers and temporally shifts evolutionary responses to periodic environments (e.g., seasonality). Stronger FD, with slope D < –2, can produce self-sustained cycles that may overwhelm responses to a changing environment, or even chaos that fundamentally limits predictability. This diversity of expected outcomes, together with the empirical evidence for both FD and environment-dependent selection, suggests that the interplay of internal dynamics with external forcing should be investigated more systematically to reach a better understanding and prediction of evolution.

We use a consumption based asset pricing model to show that the predictability of excess returns on risky assets can arise from only two sources: (1) stochastic volatility of fundamental variables, or (2) departures from rational expectations that give rise to predictable investor forecast errors and market inefficiency. While controlling for stochastic volatility, we find that a variable which measures non-fundamental noise in the Treasury yield curve helps to predict 1-month-ahead excess stock returns, but only during sample periods that include the Great Recession. For these sample periods, higher noise predicts lower excess stock returns, implying that a shortage of arbitrage capital in financial markets caused excess returns to drop below the levels justified by fundamentals. The statistical significance of the predictor variables that control for stochastic volatility are also typically sensitive to the sample period. Measures of implied and realized stock return variance cease to be significant when the COVID-influenced data from early 2020 onward is included.

In this study, we conducted an ensemble retrospective prediction from 1881 to 2017 using the Community Earth System Model to evaluate El Niño–Southern Oscillation (ENSO) predictability and its variability on different time scales. To our knowledge, this is the first assessment of ENSO predictability using a long-term ensemble hindcast with a complicated coupled general circulation model (CGCM). Our results indicate that both the dispersion component (DC) and signal component (SC) contribute to the interannual variation of ENSO predictability (measured by relative entropy). Specifically, the SC is more important for ENSO events, whereas the DC is of comparable importance for short lead times and in weak ENSO signal years. The SC dominates the seasonal variation of ENSO predictability, and an abrupt decrease in signal intensity results in the spring predictability barrier feature of ENSO. At the interdecadal scale, the SC controls the variability of ENSO predictability, while the magnitude of ENSO predictability is determined by the DC. The seasonal and interdecadal variations of ENSO predictability in the CGCM are generally consistent with results based on intermediate complexity and hybrid coupled models. However, the DC has a greater contribution in the CGCM than that in the intermediate complexity and hybrid coupled models.Significance StatementEl Niño–Southern Oscillation (ENSO) is a prominent interannual signal in the global climate system with widespread climatic influence. Our current understanding of ENSO predictability is based mainly on long-term retrospective forecasts obtained from intermediate complexity and hybrid coupled models. Compared with those models, complicated coupled general circulation models (CGCMs) include more realistic physical processes and have the potential to reproduce the ENSO complexity. However, hindcast studies based on CGCMs have only focused on the last 20–60 years. In this study, we conducted an ensemble retrospective prediction from 1881 to 2017 using the Community Earth System Model in order to evaluate ENSO predictability and examine its variability on different time scales. To our knowledge, this is the first assessment of ENSO predictability using a long-term ensemble hindcast with a CGCM.

The land surface is a potential source of climate predictability over the Northern Hemisphere midlatitudes but has received less attention than sea surface temperature in this regard. This study quantified the degree to which realistic land initialization contributes to interannual climate predictability over Europe based on a coupled climate system model named FGOALS-g2. The potential predictability provided by the initialization, which incorporates the soil moisture and soil temperature of a land surface reanalysis product into the coupled model with a dimension-reduced projection four-dimensional variational data assimilation (DRP-4DVar)-based weakly coupled data assimilation (WCDA) system, was analyzed first. The effective predictability (i.e., prediction skill) of the hindcasts by FGOALS-g2 with realistic and well-balanced initial conditions from the initialization were then evaluated. Results show an enhanced interannual prediction skill for summer surface air temperature and precipitation in the hindcast over Europe, demonstrating the potential benefit from realistic land initialization. This study highlights the significant contributions of land surface to interannual predictability of summer climate over Europe.

In this study, the predictability of the El Nino-South Oscillation (ENSO) in an operational prediction model from the perspective of initial errors is diagnosed using the seasonal hindcasts of the Beijing Climate Center System Model, BCC_CSM1.1(m). Forecast skills during the different ENSO phases are analyzed and it is shown that the ENSO forecasts appear to be more challenging during the developing phase, compared to the decay phase. During ENSO development, the SST prediction errors are significantly negative and cover a large area in the central and eastern tropical Pacific, thus limiting the model skill in predicting the intensity of El Nino. The large-scale SST errors, at their early stage, are generated gradually in terms of negative anomalies in the subsurface ocean temperature over the central-western equatorial Pacific, featuring an error evolutionary process similar to that of El Nino decay and the transition to the La Nina growth phase. Meanwhile, for short lead-time ENSO predictions, the initial wind errors begin to play an increasing role, particularly in linking with the subsurface heat content errors in the central-western Pacific. By comparing the multiple samples of initial fields in the model, it is clearly found that poor SST predictions of the Nino-3.4 region are largely due to contributions of the initial errors in certain specific locations in the tropical Pacific. This demonstrates that those sensitive areas for initial fields in ENSO prediction are fairly consistent in both previous ideal experiments and our operational predictions, indicating the need for targeted observations to further improve operational forecasts of ENSO.

International audience; Predicting how humans move within space and time is a central topic in many scientific domains such as epidemic propagation, urban planning, and ride-sharing. However, current studies neglect individuals' preferences to explore and discover new areas. Yet, neglecting novelty-seeking activities at first glance appears to be inconsequential on the ability to understand and predict individuals' trajectories. We claim and show the opposite in this work: exploration-like visits strongly impact mobility understanding and anticipation. We start by proposing a new approach to identifying exploration visits. Based on that, we construct individuals' mobility profiles using their exploration inclinations-Scouters (i.e., extreme explorers), Routiners (i.e., extreme returners), and Regulars (i.e., with no extreme behavior). Finally, we evaluate the impacts of novelty-seeking, quality of the data, and the prediction task formulation on the theoretical and practical predictability extents. The results show the validity of our profiling and highlight the obstructive impacts of noveltyseeking activities on the predictability of human trajectories. In particular, in the next-place prediction task, from 40% to 90% of predicted locations are wrong, notably with Scouters.

An accurate prediction of El Nino Southern Oscillation (ENSO) holds the key to produce skillful seasonal weather predictions across the globe. All the statistical and dynamical ENSO models developed in the past four decades face a common problem, spring predictability barrier, which is the sudden drop in the skill of the ENSO prediction when the forecast is initiated before the onset of boreal summer. Recent studies suggest that data-driven machine learning models can overcome the spring predictability barrier. We show that using a convolutional long short-term memory (ConvLSTM) network, the monthly mean Nino3.4 index can be skillfully predicted up to one year ahead. The model was also able to predict strong El Nino cases, such as 1997-1998 and 2015-2016 a year ahead. Our results suggest that the proposed ConvLSTM model has significant skill in multiseasonal weather predictions.

The predictability of geopolitical risks (GPR) for exchange rate volatility of the BRICS is examined using both historical and recent GPR data. Relying on the GARCH-MIDAS-X model based on available data frequencies, we find that the BRICS exchange rates are more vulnerable to recent GPR data than the historical data. Additional analysis suggests contrasting evidence between the recent global GPR data and the country-specific GPR data implying that the BRICS exchange rates are more vulnerable to global than domestic GPR. Finally, we document some out-of-sample economic gains of accounting for GPR in the valuation of foreign exchange portfolio.

In anticipating upcoming content, comprehenders are known to rely on real-world knowledge. This knowledge can be deployed directly in favor of upcoming content about typical situations (implying a transparent mapping between the world and what speakers say about the world). Such knowledge can also be used to estimate the likelihood of speech, whereby atypical situations are the ones newsworthy enough to merit reporting (i.e., a nontransparent mapping in which improbable situations yield likely utterances). We report four forced-choice studies (three preregistered) testing this distinction between situation knowledge and speech production likelihood. Comprehenders are shown to anticipate situation-atypical meanings more when guessing content (a) that a speaker announces (rather than thinks), (b) that is said out of the blue (rather than produced when prompted), and (c) that is addressed to a large audience (rather than a single listener). The findings contrast with prior work that emphasizes a comprehension bias in favor of typicality, and they highlight the need for comprehension models that incorporate expectations for informativity (as one of a set of inferred speaker goals) alongside expectations for content plausibility.

I find no evidence that partial least squares based on disaggregated book-to-market ratios produces a model of market premiums with persistently positive out-of-sample R2, as originally documented for market returns. This is consistent with time variation in predictability, for example, and does not necessarily invalidate the method. The two main drivers of the original performance are: (i) The sample period and (ii) the use of market returns as forecasting targets. Two other drivers are using, as regressors, (iii) the book-to-market ratios of the specific portfolios double-sorted by size and book-to-market (iv) divided by their standard deviations.

We investigate the sensitivity of mesoscale atmospheric predictability to the slope of the background kinetic energy spectrum E by adding initial errors to simulations of idealized moist midlatitude cyclones at several wavenumbers k for which the slope of E(k) is significantly different. These different slopes arise from 1) differences in the E(k) generated by cyclones growing in two different moist baroclinically unstable environments, and 2) differences in the horizontal scale at which initial perturbations are added, with E(k) having steeper slopes at larger scales. When small-amplitude potential temperature perturbations are added, the error growth through the subsequent 36-h simulation is not sensitive to the slope of E(k) nor to the horizontal scale of the initial error. In all cases with small-amplitude perturbations, the error growth in physical space is dominated by moist convection along frontal boundaries. As such, the error field is localized in physical space and broad in wavenumber (spectral) space. In moist midlatitude cyclones, these broadly distributed errors in wavenumber space limit mesoscale predictability by growing up-amplitude rather than by cascading upscale to progressively longer wavelengths. In contrast, the error distribution in homogeneous turbulence is broad in physical space and localized in wavenumber space, and dimensional analysis can be used to estimate the error growth rate at a specific wavenumber k from E(k). Predictability estimates derived in this manner, and from the numerical solutions of idealized models of homogeneous turbulence, depend on whether the slope of E(k) is shallower or steeper than k−3, which differs from the slope-insensitive behavior exhibited by moist midlatitude cyclones.

The onset of the rainy season is an important date for the mostly rain-fed agricultural practices in Vietnam. Subseasonal to seasonal (S2S) ensemble hindcasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) are used to evaluate the predictability of the rainy season onset dates (RSODs) over five climatic subregions of Vietnam. The results show that the ECMWF model reproduces well the observed interannual variability of RSODs, with a high correlation ranging from 0.60 to 0.99 over all subregions at all lead times (up to 40 days) using five different RSOD definitions. For increasing lead times, forecasted RSODs tend to be earlier than the observed ones. Positive skill score values for almost all cases examined in all subregions indicate that the model outperforms the observed climatology in predicting the RSOD at subseasonal lead times (∼28–35 days). However, the model is overall more skillful at shorter lead times. The choice of the RSOD criterion should be considered because it can significantly influence the model performance. The result of analyzing the highest skill score for each subregion at each lead time shows that criteria with higher 5-day rainfall thresholds tend to be more suitable for the forecasts at long lead times. However, the values of mean absolute error are approximately the same as the absolute values of the mean error, indicating that the prediction could be improved by a simple bias correction. The present study shows a large potential to use S2S forecasts to provide meaningful predictions of RSODs for farmers.

We develop a new housing search index (HSI) extracted from online search activity on a limited set of keywords related to the house-buying process. We show that HSI has strong predictive power over subsequent changes in house prices, both in-sample and out-of-sample and after controlling for the effect of commonly used predictors, and relate our findings to models of search-induced frictions. Our results imply that search data can be used as an early indicator of where the market is going. This paper was accepted by David Sraer, finance. Funding: We acknowledge support from the Danish Finance Institute and from the Independent Research Fund Denmark [Grants DFF 7024-00020B and DFF 7015-00017B]. Supplemental Material: The online appendix and data are available at https://doi.org/10.1287/mnsc.2023.4672 .

In June 2021, Western North America experienced an intense heat wave with unprecedented temperatures and far-reaching socio-economic consequences. Anomalous rainfall in the West Pacific triggers a cascade of weather events across the Pacific, which build up a high-amplitude ridge over Canada and ultimately lead to the heat wave. We show that the response of the jet stream to diabatically enhanced ascending motion in extratropical cyclones represents a predictability barrier with regard to the heat wave magnitude. Therefore, probabilistic weather forecasts are only able to predict the extremity of the heat wave once the complex cascade of weather events is captured. Our results highlight the key role of the sequence of individual weather events in limiting the predictability of this extreme event. We therefore conclude that it is not sufficient to consider such rare events in isolation but it is essential to account for the whole cascade over different spatiotemporal scales.

Contemporary evolution has the potential to significantly alter biotic responses to global change, including range expansion dynamics and biological invasions. However, predictive models often make highly simplifying assumptions about the genetic architecture underlying relevant traits. This can be problematic since genetic architecture defines evolvability, that is, evolutionary rates, and higher order evolutionary processes, which determine whether evolution will be able to keep up with environmental change or not. Therefore, we here study the impact of the genetic architecture of dispersal and local adaptation, two central traits of high relevance for range expansion dynamics, on the speed and variability of range expansions into an environmental gradient, such as temperature. In our theoretical model we assume that dispersal and local adaptation traits result from the products of two non-interacting gene-regulatory networks (GRNs). We compare our model to simpler quantitative genetics models and show that in the GRN model, range expansions are accelerated, faster and more variable. Increased variability implies that these evolutionary changes reduce predictability. We further find that acceleration in the GRN model is primarily driven by an increase in the rate of local adaptation to novel habitats which results from greater sensitivity to mutation (decreased robustness) and increased gene expression. Our results highlight how processes at microscopic scales, here, within genomes, can impact the predictions of large scale, macroscopic phenomena, such as range expansions, by modulating the rate of evolution.