Search

We study the performance of a sequential linear prediction system built on nanoscale beyond-CMOS circuit fabric that may introduce in computation. We propose a new sequential linear prediction algorithm under a mixture-of-experts framework that performs satisfactorily in the presence of computational errors. We introduce a worst-case approach to model the computational errors, where we view erroneous circuit fabric as an adversary that perturbs the prediction algorithm to heavily deteriorate its performance. We demonstrate that our algorithm achieves uniformly good performance under the worst-case error approach in an individual sequence manner.

Procedural content generation is a popular technique in the game development. One of its typical applications is generation of game levels. This paper presents a method to generate tracks for racing games, by viewing racing track generation as a discrete sequence prediction problem. To solve it we combine two techniques from music improvisation. We show that this method is capable of generating new racing tracks which appear to be interesting.

SummaryObjectives: The goal of smart homes is to create an intelligent environment adapting the inhabitants need and assisting the person who needs special care and safety in their daily life. This can be reached by collecting the ADL (activities of daily living) data and further analysis within existing computing elements. In this research, a very recent algorithm named sequence prediction via enhanced episode discovery (SPEED) is modified and in order to improve accuracy time component is included.Methods: The modified SPEED or M-SPEED is a sequence prediction algorithm, which modified the previous SPEED algorithm by using time duration of appliance’s ON-OFF states to decide the next state. M-SPEED discovered periodic episodes of inhabitant behavior, trained it with learned episodes, and made decisions based on the obtained knowledge.Results: The results showed that M-SPEED achieves 96.8% prediction accuracy, which is better than other time prediction algorithms like PUBS, ALZ with temporal rules and the previous SPEED.Conclusions: Since human behavior shows natural temporal patterns, duration times can be used to predict future events more accurately. This inhabitant activity prediction system will certainly improve the smart homes by ensuring safety and better care for elderly and handicapped people.

In our paper, we substantiate the fact that codon 104 variation is adaptive in highland Tibet plateau mammals (1). Liu’s letter (2) claims that codon 104 is not adaptive, which is only based on searching the sequence prediction in National Center for Biotechnology Information (NCBI) without experimental evidence.

The field of computational protein design is reaching its adolescence. Protein design algorithms have been applied to design or engineer proteins that fold, fold faster, catalyze, catalyze faster, signal, and adopt preferred conformational states. Further developments of scoring functions, sampling strategies, and optimization methods will expand the range of applicability of computational protein design to larger and more varied systems, with greater incidence of success. Developments in this field are beginning to have significant impact on biotechnology and chemical biology.

Motivation: How critical is the sequence order information in predicting protein secondary structure segments? We tried to get a rough insight on it from a theoretical approach using both a prediction algorithm and structural fragments from Protein Databank (PDB). Results: Using reverse protein sequences and PDB structural fragments, we theoretically estimated the significance of the order for protein secondary structure and prediction. On average: (1) 79% of protein sequence segments resulted in the same prediction in both normal and reverse directions, which indicated a relatively high conservation of secondary structure propensity in the reverse direction; (2) the reversed sequence prediction alone performed less accurately than the normal forward sequence prediction, but comparably high (2% difference); (3) the commonly predicted regions showed a slightly higher prediction accuracy (4%) than the normal sequences prediction; and (4) structural fragments which have counterparts in reverse direction in the same protein showed a comparable degree of secondary structure conservation (73% identity with reversed structures on average for pentamers). Contact: jong@biosophy.org; dietmann@ebi.ac.uk; heger@ebi.ac.uk; holm@ebi.ac.uk *** Present address: EBI, Genome Campus, Hinxton, Cambridge, CB10 1SD, UK. To whom correspondence should be addressed.

For any discrete-state sequence prediction algorithm A, it is always possible, using an algorithm B no more complicated than A, to generate a sequence for which A's prediction is always wrong. For any prediction algorithm A and sequence x, there exists a sequence y no more complicated than x, such that if A performs better than random on x, then it will perform worse than random on y by the same margin. An example of a simple neural network predicting a bit sequence is used to illustrate this very general but not widely recognized phenomenon. This implies that any predictor with good performance must rely on some (usually implicitly) assumed prior distributions of the problem.

The goal of this work was to develop sequential prediction methods for online information fusion and control, with methods designed to handle unknown, environmental dynamics, potentially stemming from an adversary who reacts to sensing actions, active sensing paradigms, and external feedback mechanisms. Online prediction and targeted collection of information is an emerging paradigm at the intersection of optimization, machine learning and control theory, which is concerned with real-time sequential planning of actions or decisions in the presence of model uncertainty, nonstationarity, and possibly adversarial disturbances. Several methods and underlying supporting theory which meet these objectives are described in detail in the following final report.

This paper investigates how noise affects a minimal computational model of the hippocampus and, in particular, region CA3. The architecture and physiology employed are consistent with the known anatomy and physiology of this region. Here, we use computer simulations to demonstrate and quantify the ability of this model to create context codes in sequential learning problems. These context codes are mediated by local context neurons which are analogous to hippocampal place-coding cells. These local context neurons endow the network with many of its problem-solving abilities. Our results show that the network encodes context on its own and then uses context to solve sequence prediction under ambiguous conditions. Noise during learning affects performance, and it also affects the development of context codes. The relationship between noise and performance in a sequence prediction is simple and corresponds to a disruption of local context neuron firing. As noise exceeds the signal, sequence completion and local context neuron firing are both lost. For the parameters investigated, extra learning trials and slower learning rates do not overcome either of the effects of noise. The results are consistent with the important role played, in this hippocampal model, by local context neurons in sequence prediction and for disambiguation across time.

The splice sites are essential for pre-mRNA maturation and crucial for Splice Site Modelling (SSM); however, there are gaps between the splicing signals and the computationally identified sequence features. In this paper, the Locality Sensitive Features (LSFs) are proposed to reduce the gaps by homogenising their contexts. Under the skewness-kurtosis based statistics and data analysis, SSM attributed with LSFs is fulfilled by double-boundary outlier filters. The LSF-based SSM had been applied to six model organisms of diverse species; by the accuracy and Receiver Operating Characteristic (ROC) analysis, the promising results show the proposed methodology is versatile and robust for the splice-site classification. It is prospective the LSF-based SSM can serve as a new infrastructure for developing effective splice-site prediction methods and have the potential to be applied to other sequence prediction problems.

We investigate behavioral prediction approaches based on subspace methods such as principal component analysis (PCA) and independent component analysis (ICA). Moreover, we propose a personalized sequential prediction approach to predict next day behavior based on features extracted from past behavioral data using subspace methods. The proposed approach is applied to the individual call (voice calls and short messages) behavior prediction task. Experimental results on the Nokia mobility data challenge (MDC) dataset are used to show the feasibility of our proposed prediction approach. Furthermore, we investigate whether prediction accuracy can be improved (i) when specific call type (voice call or short message), instead of the general call behavior prediction, is considered in the prediction task, and (ii) when workday and weekend scenarios are considered separately.

We discuss algorithms for combining sequential prediction strategies, a task which can be viewed as a natural generalization of the concept of universal coding. We describe a graphical language based on hidden Markov models for defining prediction strategies, and we provide both existing and new models as examples. The models include efficient, parameterless models for switching between the input strategies over time, including a model for the case where switches tend to occur in clusters, and finally a new model for the scenario where the prediction strategies have a known relationship, and where jumps are typically between strongly related ones. This last model is relevant for coding time series data where parameter drift is expected. As theoretical contributions, we introduce an interpolation construction that is useful in the development and analysis of new algorithms, and we establish a new sophisticated lemma for analyzing the individual sequence regret of parameterized models.

Given a growing sequence of observations x_1,...,x_n,..., one is required, at each time step n, to make some inference about the stochastic mechanism generating the sequence. Several problems that have numerous applications in different branches of mathematics and computer science can be formulated in this way. For example, one may want to forecast probabilities of the next outcome x_{n+1} (sequence prediction); to make a decision on whether the mechanism generating the sequence belongs to a certain family $H_0$ versus it belongs to a different family $H_1$ (hypothesis testing); to take an action in order to maximize some utility function. In each of these problems, as well as in many others, in order to be able to make inference, one has to make some assumptions on the probabilistic mechanism generating the data. Typical assumptions are that x_i are independent and identically distributed, or that the distribution generating the sequence belongs to a certain parametric family. The central question addressed in this work is: under which assumptions is inference possible? This question is considered for several problems of inference, including sequence prediction, hypothesis testing, classification and reinforcement learning.; Les travaux présentés sont dédiés à la possibilité de faire de l'inférence statistique à partir de données séquentielles. Le problème est le suivant. Étant donnée une suite d'observations x_1,...,x_n,..., on cherche à faire de l'inférence sur le processus aléatoire ayant produit la suite. Plusieurs problèmes, qui d'ailleurs ont des applications multiples dans différents domaines des mathématiques et de l'informatique, peuvent être formulés ainsi. Par exemple, on peut vouloir prédire la probabilité d'apparition de l'observation suivante, x_{n+1} (le problème de prédiction séquentielle); ou répondre à la question de savoir si le processus aléatoire qui produit la suite appartient à un certain ensemble H_0 versus appartient à un ensemble différent H_1 (test d'hypothèse) ; ou encore, effectuer une action avec le but de maximiser une certain fonction d'utilité. Dans chacun de ces problèmes, pour rendre l'inférence possible il faut d'abord faire certaines hypothèses sur le processus aléatoire qui produit les données. La question centrale adressée dans les travaux présentés est la suivante : sous quelles hypothèses l'inférence est-elle possible ? Cette question est posée et analysée pour des problèmes d'inférence différents, parmi lesquels se trouvent la prédiction séquentielle, les tests d'hypothèse, la classification et l'apprentissage par renforcement.

Given a growing sequence of observations x_1,...,x_n,..., one is required, at each time step n, to make some inference about the stochastic mechanism generating the sequence. Several problems that have numerous applications in different branches of mathematics and computer science can be formulated in this way. For example, one may want to forecast probabilities of the next outcome x_{n+1} (sequence prediction); to make a decision on whether the mechanism generating the sequence belongs to a certain family $H_0$ versus it belongs to a different family $H_1$ (hypothesis testing); to take an action in order to maximize some utility function. In each of these problems, as well as in many others, in order to be able to make inference, one has to make some assumptions on the probabilistic mechanism generating the data. Typical assumptions are that x_i are independent and identically distributed, or that the distribution generating the sequence belongs to a certain parametric family. The central question addressed in this work is: under which assumptions is inference possible? This question is considered for several problems of inference, including sequence prediction, hypothesis testing, classification and reinforcement learning.; Les travaux présentés sont dédiés à la possibilité de faire de l'inférence statistique à partir de données séquentielles. Le problème est le suivant. Étant donnée une suite d'observations x_1,...,x_n,..., on cherche à faire de l'inférence sur le processus aléatoire ayant produit la suite. Plusieurs problèmes, qui d'ailleurs ont des applications multiples dans différents domaines des mathématiques et de l'informatique, peuvent être formulés ainsi. Par exemple, on peut vouloir prédire la probabilité d'apparition de l'observation suivante, x_{n+1} (le problème de prédiction séquentielle); ou répondre à la question de savoir si le processus aléatoire qui produit la suite appartient à un certain ensemble H_0 versus appartient à un ensemble différent H_1 (test d'hypothèse) ; ou encore, effectuer une action avec le but de maximiser une certain fonction d'utilité. Dans chacun de ces problèmes, pour rendre l'inférence possible il faut d'abord faire certaines hypothèses sur le processus aléatoire qui produit les données. La question centrale adressée dans les travaux présentés est la suivante : sous quelles hypothèses l'inférence est-elle possible ? Cette question est posée et analysée pour des problèmes d'inférence différents, parmi lesquels se trouvent la prédiction séquentielle, les tests d'hypothèse, la classification et l'apprentissage par renforcement.

Neurons in many brain regions change their spiking responses and interactions among them to relevant stimuli. Tracking the dynamics of neural system is crucial for understanding how neural systems adapt their responses to relevant biological information. Granger causality [1] has been effectively used to assess directional interactions between continuous neural signals, but it cannot be directly applied to neural spike trains viewed as point processes. Recently, methods that extend Granger’s viewpoint to the point process modality have been developed [2], [3] to identify causal interactions between neural spike trains. These methods, however, depend upon stationarity assumptions – which might not be valid when the interactive causal influences themselves are time-varying. We propose a novel probabilistic method for tracking the time-varying causal neural interactions based on sequential prediction of point process models. The time-varying causality from neuron x to y is assessed by the variability of a windowed version of the point process log-likelihood ratio (LLR), where one model incorporates only the past y and the other incorporates the past of both x and y. The proposed method successfully tracks the time-varying causal network in simulated data, and when applied to real neural data recorded in the rat insular cortex, it identifies the change of causal relationships between neurons to a relevant behavioral stimulus (see Figure ​Figure11). Figure 1 Tracking time-varying causality network. Sij(t) represents time-varying causality effect from neuron i to j. A. Simulation: Proposed time-varying causality measure had larger values when x caused y than when x did not cause y. B. Real data analysis: Time-varying ...

In his 1969 paper, Granger proposed a statistical definition of causality between stochastic processes. It is based on whether causal side information helps in a sequential prediction task. However, his formulation was limited to linear predictors. We describe a generalized framework, where predictions are beliefs and compare the best predictor with side information to the best predictor without side information. The difference in the prediction performance, i.e., regret of such predictors, is used as a measure of causal influence of the side information. Specifically when log loss is used to quantify each predictor's loss and an expectation over the outcomes is used to quantify the regret, we show that the directed information, an information theoretic quantity, quantifies Granger causality. We also explore a more pessimistic setup perhaps better suited for adversarial settings where minimax criterion is used to quantify the regret.

We study the problem of finding an universal estimation scheme $h_n:\mathbb{R}^n\to \mathbb{R}$, $n=1,2,...$ which will satisfy \lim_{t\rightarrow\infty}{\frac{1}{t}}\sum_{i=1}^t|h_ i(X_0,X_1,...,X_{i-1})-E(X_i|X_0,X_1,...,X_{i-1})|^p=0 a.s. for all real valued stationary and ergodic processes that are in $L^p$. We will construct a single such scheme for all $1, Comment: Published in at http://dx.doi.org/10.1214/10-AOP576 the Annals of Probability (http://www.imstat.org/aop/) by the Institute of Mathematical Statistics (http://www.imstat.org)

We study sequential prediction models in cases where only fragments of the sequences are annotated with the ground-truth. The task does not match the standard semi-supervised setting and is highly relevant in areas such as natural language processing, where completely labeled instances are expensive and require editorial data. We propose to generalize the semi-supervised setting and devise a simple transductive loss-augmented perceptron to learn from inexpensive partially annotated sequences that could for instance be provided by laymen, the wisdom of the crowd, or even automatically. Experiments on mono- and crosslingual named entity recognition tasks with automatically generated partially annotated sentences from Wikipedia demonstrate the effectiveness of the proposed approach. Our results show that learning from partially labeled data is never worse than standard supervised and semi-supervised approaches trained on data with the same ratio of labeled and unlabeled tokens.

Background: The structure conservation in various ?-helix subclasses reveals the sequence and context dependent factors causing distortions in the ?-helix. The sequence-structure relationship in these subclasses can be used to predict structural variations in ?-helix purely based on its sequence. We train support vector machine(SVM) with dot product kernel function to discriminate between regular ?-helix and non-regular ?-helices purely based on the sequences, which are represented with various overall and position specific propensities of amino acids.Results: We characterize the structural distortions in five ?-helix subclasses. The sequence structure correlation in the subclasses reveals that the increased propensity of proline, histidine, serine, aspartic acid and aromatic amino acids are responsible for the distortions in regular ?-helix. The N-terminus of regular ?-helix prefers neutral and acidic polar amino acids, while the C-terminus prefers basic polar amino acid. Proline is preferred in the first turn of regular ?-helix , while it is preferred to produce kinked and curved subclasses. The SVM discriminates between regular ?-helix and the rest with precision of 80.97% and recall of 88.05%.Conclusions: The correlation between structural variation in helices and their sequences is manifested by the performance of SVM based on sequence features. The results presented here are useful for computational design of helices. The results are also useful for prediction of structural perturbations in helix sequence purely based on its sequence. � 2011 Tendulkar and Wangikar; licensee BioMed Central Ltd.

What is the relationship between the complexity of a learner and the randomness of his mistakes ? This question was posed in [4] who showed that the more complex the learner the higher the possibility that his mistakes deviate from a true random sequence. In the current paper we report on an empirical investigation of this problem. We investigate two characteristics of randomness, the stochastic and algorithmic complexity of the binary sequence of mistakes. A learner with a Markov model of order k is trained on a finite binary sequence produced by a Markov source of order k ∗ and is tested on a different random sequence. As a measure of learner’s complexity we define a quantity called the sysRatio, denoted by ρ, which is the ratio between the compressed and uncompressed lengths of the binary string whose i th bit represents the maximum a posteriori decision made at state i of the learner’s model. The quantity ρ is a measure of information density. The main result of the paper shows that this ratio is crucial in answering the above posed question. The result indicates that there is a critical threshold ρ ∗ such that when ρ ≤ ρ ∗ the sequence of mistakes possesses the following features: (1) low divergence ∆ from a random sequence, (2) low variance in algorithmic complexity. When ρ > ρ∗ , the characteristics of the mistake sequence changes sharply towards a high ∆ and high variance in algorithmic complexity. It is also shown that the quantity ρ is inversely proportional to k and the value of ρ ∗ corresponds to the value k ∗ . This is the point where the learner’s model becomes too simple and is unable to approximate the Bayes optimal decision. Here the characteristics of the mistake sequence change sharply, https://www.edusoft.ro/brain/index.php/brain/article/view/114/241

International audience; A sequence $x_1,\dots,x_n,\dots$ of discrete-valued observations is generated according to some unknown probabilistic law (measure) $\mu$. After observing each outcome, it is required to give the conditional probabilities of the next observation. The realizable case is when the measure $\mu$ belongs to an arbitrary but known class $C$ of process measures. The non-realizable case is when $\mu$ is completely arbitrary, but the prediction performance is measured with respect to a given set $C$ of process measures. We are interested in the relations between these problems and between their solutions, as well as in characterizing the cases when a solution exists, and finding these solutions. We show that if the quality of prediction is measured by total variation distance, then these problems coincide, while if it is measured by expected average KL divergence, then they are different. For some of the formalizations we also show that when a solution exists, it can be obtained as a Bayes mixture over a countable subset of $C$. As an illustration to the general results obtained, we show that a solution to the non-realizable case of the sequence prediction problem exists for the set of all finite-memory processes, but does not exist for the set of all stationary processes.

Time series prediction covers a vast field of every-day statistical applications in medical, environmental and economic domains. In this paper we develop nonparametric prediction strategies based on the combination of a set of 'experts' and show the universal consistency of these strategies under a minimum of conditions. We perform an in-depth analysis of real-world data sets and show that these nonparametric strategies are more flexible, faster and generally outperform ARMA methods in terms of normalized cumulative prediction error., Comment: article + 2 figures

With combination of grey neural network (CGNN) and grey relational clustering, the models are constructed, which are used to solve the prediction and comparison of surrounding rocks stability controlling parameters in deep entry of coal mine. The results show that grey relational clustering is an effective way and CGNN has perfect ability to be studied in a short-term prediction. Combined grey neural network has the features of trend and fluctuation while combining with the time-dependent sequence prediction. It is concluded that great improvements compared with any methods of trend prediction and simple factor in combined grey neural network is stated and described in stably controlling the surrounding rocks in deep entry.

Using grey relation clustering and combined grey neural network (CGNN), the combined model is setup, which aims at solving the problems of predicting and comparing the gas concentration in top corner of coal mine. Through comparison and prediction, the results show that, in short-term prediction, grey relation clustering is an effective way and CGNN has perfect ability to study. CGNN has the dual properties of trend and fluctuation under the condition of combining with the time-dependent sequence prediction. It is concluded that great improvement comparing with any methods of trend prediction and simple factor in CGNN is stated and described in gas concentration in top corner of coal mine.

A Gray Model for international tourists prediction in China is established based on Gray System Theory and related data since 1978. Tested by Posterior Check and Small Probability, the model is proved qualified to predict the number of international tourists with high elevation fitting precision. Gray System theory is a powerful tool to study tourist phenomena, because tourism system is belonging to Gray System. This paper will be a useful reference to the study on tourism sequence prediction.

This paper investigates the implementation of an Artificial Neural Network (ANN) model for sequential prediction of daily groundwater levels based on precipitation forecasts. The basic principle of the ANN-based procedure consists of relating previous daily groundwater levels and daily precipitation forecasts in order to predict daily groundwater levels up to seven days ahead. The daily precipitation values up to one week ahead are assumed to be deterministic since meteorological short-range forecasts are generally available. The methodology is applied to the groundwater system of Matsuyama City, Japan. Insufficiency of water is a periodical problem in this city and thus accurate predictions of groundwater levels are very important to improve the water resources management in the region. The excellent accuracy obtained by the ANN model indicates that it is very efficient for the multi-step-ahead prediction of daily groundwater levels. As conclusion, this methodology may provide trustworthy data for the application of models to the sustainable management of Matsuyama’s groundwater system.

We present DirecTL: an online discriminative sequence prediction model that employs a many-to-many alignment between target and source. Our system incorporates input segmentation, target character prediction, and sequence modeling in a unified dynamic programming framework. Experimental results suggest that DirecTL is able to independently discover many of the language-specific regularities in the training data.

International audience; The problem is sequence prediction in the following setting. A sequence $x_1,\dots,x_n,\dots$ of discrete-valued observations is generated according to some unknown probabilistic law (measure) $\mu$. After observing each outcome, it is required to give the conditional probabilities of the next observation. The measure $\mu$ belongs to an arbitrary but known class $C$ of stochastic process measures. We are interested in predictors $\rho$ whose conditional probabilities converge (in some sense) to the ``true'' $\mu$-conditional probabilities if any $\mu\in C$ is chosen to generate the sequence. The contribution of this work is in characterizing the families $C$ for which such predictors exist, and in providing a specific and simple form in which to look for a solution. We show that if any predictor works, then there exists a Bayesian predictor, whose prior is discrete, and which works too. We also find several sufficient and necessary conditions for the existence of a predictor, in terms of topological characterizations of the family $C$, as well as in terms of local behaviour of the measures in $C$, which in some cases lead to procedures for constructing such predictors. It should be emphasized that the framework is completely general: the stochastic processes considered are not required to be i.i.d., stationary, or to belong to any parametric or countable family.

In this paper the sequential prediction problem with expert advice is considered for the case when the losses of experts suffered at each step can be unbounded. We present some modification of Kalai and Vempala algorithm of following the perturbed leader where weights depend on past losses of the experts. New notions of a volume and a scaled fluctuation of a game are introduced. We present an algorithm protected from unrestrictedly large one-step losses. This algorithm has the optimal performance in the case when the scaled fluctuations of onestep losses of experts of the pool tend to zero.

MicroRNAs (miRNAs) are single-stranded non-coding RNAs shown to plays important regulatory roles in a wide range of biological processes and diseases. The functions and regulatory mechanisms of most of miRNAs are still poorly understood in part because of the difficulty in identifying the miRNA regulatory targets. To this end, computational methods have evolved as important tools for genome-wide target screening. Although considerable work in the past few years has produced many target prediction algorithms, most of them are solely based on sequence, and their accuracy is still poor. In contrast, gene expression profiling from miRNA over-expression experiments can provide additional information about miRNA targets. This paper presents a Bayesian approach to integrate sequence level prediction result with expression profiling to improve the performance of miRNA target identification. The test on proteomic and IP pull-down data demonstrated better performance of the proposed approach.

We flrst deflne Algorithmic Probability, an extremely powerful method of inductive inference. We discuss its completeness, incomputability, diversity and subjectivity and show that its incomputability in no way inhibits its use for practical prediction. Applications to Bernoulli sequence prediction and grammar discovery are described. We conclude with a note on its employment in a very strong AI system for very general problem solving.

We consider the problem of sequence prediction in a probabilistic setting. Let there be given a class $\mathcal C$ of stochastic processes (probability measures on the set of one-way infinite sequences). We are interested in the question of what are the conditions on $\mathcal C$ under which there exists a predictor (also a stochastic process) for which the predicted probabilities converge to the correct ones if any of the processes in $\mathcal C$ is chosen to generate the data. We find some sufficient conditions on $\mathcal C$ under which such a predictor exists. Some of the conditions are asymptotic in nature, while others are based on the local (truncated to first observations) behaviour of the processes. The conditions lead to constructions of the predictors. In some cases we obtain rates of convergence that are optimal up to an additive logarithmic term. We emphasize that the framework is completely general: the stochastic processes considered are not required to be i.i.d., stationary, or to belong to some parametric family.

International audience; Suppose we are given two probability measures on the set of one-way infinite finite-alphabet sequences. Consider the question of when one of the measures predicts the other, that is, when conditional probabilities converge (in a certain sense), if one of the measures is chosen to generate the sequence. This question may be considered a refinement of the problem of sequence prediction in its most general formulation: for a given class of probability measures, does there exist a measure which predicts all of the measures in the class? To address this problem, we find some conditions on local absolute continuity which are sufficient for prediction and generalize several different notions that are known to be sufficient for prediction. We also formulate some open questions to outline a direction for finding the conditions on classes of measures for which prediction is possible.

Background A number of sequence-based methods exist for protein secondary structure prediction. Protein secondary structures can also be determined experimentally from circular dichroism, and infrared spectroscopic data using empirical analysis methods. It has been proposed that comparable accuracy can be obtained from sequence-based predictions as from these biophysical measurements. Here we have examined the secondary structure determination accuracies of sequence prediction methods with the empirically determined values from the spectroscopic data on datasets of proteins for which both crystal structures and spectroscopic data are available. Results In this study we show that the sequence prediction methods have accuracies nearly comparable to those of spectroscopic methods. However, we also demonstrate that combining the spectroscopic and sequences techniques produces significant overall improvements in secondary structure determinations. In addition, combining the extra information content available from synchrotron radiation circular dichroism data with sequence methods also shows improvements. Conclusion Combining sequence prediction with experimentally determined spectroscopic methods for protein secondary structure content significantly enhances the accuracy of the overall results obtained.

Based on record data, prediction of the future records from the two-parameter Weibull distribution is studied. First we consider the sampling based procedure to compute the Bayes estimates and also to construct symmetric credible intervals. Secondly, we consider one-sequence and two-sequence Bayes prediction of the future records based on some observed records. The Monte Carlo algorithms are used to compute simulation consistent predictors and prediction intervals for future unobserved records. A numerical simulation study is conducted to compare the different methods and a real data set involving the annual rainfall recorded at Los Angeles Civic Center during 132 years is analyzed to illustrate the procedures developed here.

The Bayesian framework is a well-studied and successful framework for inductive reasoning, which includes hypothesis testing and confirmation, parameter estimation, sequence prediction, classification, and regression. But standard statistical guidelines for choosing the model class and prior are not always available or fail, in particular in complex situations. Solomonoff completed the Bayesian framework by providing a rigorous, unique, formal, and universal choice for the model class and the prior. We discuss in breadth how and in which sense universal (non-i.i.d.) sequence prediction solves various (philosophical) problems of traditional Bayesian sequence prediction. We show that Solomonoff's model possesses many desirable properties: Strong total and weak instantaneous bounds, and in contrast to most classical continuous prior densities has no zero p(oste)rior problem, i.e. can confirm universal hypotheses, is reparametrization and regrouping invariant, and avoids the old-evidence and updating problem. It even performs well (actually better) in non-computable environments., Comment: 24 pages

This paper reports a sequential prediction system of "large scale database-based online modeling (LOM)". The sequential prediction system predicts time-series process variables repeating processing that predicts process variables of next step by using the predicted process variables of previous step and prepared manipulated variables. Furthermore, the system is applied to the industrial reactor; practical effectiveness of the system is verified. As the result, the system has predicted the process variables with satisfactory accuracy. The practical effectiveness has been confirmed.

We propose simple randomized strategies for sequential prediction under imperfect monitoring, that is, when the forecaster does not have access to the past outcomes but rather to a feedback signal. The proposed strategies are consistent in the sense that they achieve, asymptotically, the best possible average reward. It was Rustichini [11] who first proved the existence of such consistent predictors. The forecasters presented here offer the first constructive proof of consistency. Moreover, the proposed algorithms are computationally efficient. We also establish upper bounds for the rates of convergence. In the case of deterministic feedback, these rates are optimal up to logarithmic terms.

In this paper the sequential prediction problem with expert advice is considered when the loss is unbounded under partial monitoring scenarios. We deal with a wide class of the partial monitoring problems: the combination of the label efficient and multi-armed bandit problem, that is, where the algorithm is only informed about the performance of the chosen expert with probability e≤1. For bounded losses an algorithm is given whose expected regret scales with the square root of the loss of the best expert. For unbounded losses we prove that Hannan consistency can be achieved, depending on the growth rate of the average squared losses of the experts.

We address the problem of sequence prediction for nonstationary stochastic processes. In particular, given two measures on the set of one-way infinite sequences over a finite alphabet, consider the question whether one of the measures predicts the other. We find some conditions on local absolute continuity under which prediction is possible.

We outline a method for quantifying the error of a sequence prediction. With sequence predictions represented by semimeasures $ u(x)$ we define their error to be $-log_2 u(x)$. We note that enumerable semimeasures are those which model the sequence as the output of a computable system given unknown input. Using this we define the simulation complexity of a computable system $C$ relative to another $U$ giving an emph{exact} bound on their difference in error. This error in turn gives an exact upper bound on the number of predictions $ u$ gets incorrect.

Solomonoff's inductive learning model is a powerful, universal and highly elegant theory of sequence prediction. Its critical flaw is that it is incomputable and thus cannot be used in practice. It is sometimes suggested that it may still be useful to help guide the development of very general and powerful theories of prediction which are computable. In this paper it is shown that although powerful algorithms exist, they are necessarily highly complex. This alone makes their theoretical analysis problematic, however it is further shown that beyond a moderate level of complexity the analysis runs into the deeper problem of Godel incompleteness. This limits the power of mathematics to analyse and study prediction algorithms, and indeed intelligent systems in general.

No Abstract.Journal of Mining and Geology 2005, Vol. 41(1): 57-79

Robust sequence prediction is an essential component of an intelligent agent acting in a dynamic world. We consider the case of near-future event prediction by an online learning agent operating in a non-stationary environment. The challenge for a learning agent under these conditions is to exploit the relevant experience from a limited environmental event history while preserving flexibility.We propose a novel time/space efficient method for learning temporal sequences and making short-term predictions. Our method operates on-line, requires few exemplars, and adapts easily and quickly to changes in the underlying stochastic world model. Using a short-term memory of recent observations, the method maintains a dynamic space of candidate hypotheses in which the growth of the space is systematically and dynamically pruned using an entropy measure over the observed predictive quality of each candidate hypothesis.The method compares well against Markov-chain predictions, and adapts faster than learned Markov-chain models to changes in the underlying distribution of events. We demonstrate the method using both synthetic data and empirical experience from a game-playing scenario with human opponents.

This paper shows how the newly developed paradigm of General Fuzzy Automata (GFA) can be used as a biological sequence predictor. We consider the positional correlations of amino acids in a protein family as the basic criteria for prediction and classification of unknown sequences. It will be shown how the GFA formalism can be used as an efficient tool for classification of protein sequences. The results show that this approach predicts the membership of an unknown sequence in a protein family better than profile Hidden Markov Models (HMMs) which are now a popular and putative approach in biological sequence analysis.