Razor Principle Research Articles

An active human role is essential in big data-led decisions and data-intensive science.

Big data is transforming many sectors, with far-reaching consequences to how decisions are made and how knowledge is produced and shared. In the current move toward more data-led decisions and data-intensive science, we aim here to examine three issues that are changing the way data are read and used. First, there is a shift toward paradigms that involve a large amount of data. In such paradigms, the creation of complex data-led models becomes tractable and appealing to generate predictions and explanations. This necessitates for instance a rethinking of Occam's razor principle in the context of knowledge discovery. Second, there is a growing erosion of the human role in decision making and knowledge discovery processes. Human users' involvement is decreasing at an alarming rate, with no say on how to read, process, and summarize data. This makes legal responsibility and accountability hard to define. Third, thanks to its increasing popularity, big data is gaining a seductive allure, where volume and complexity of big data can de facto confer more persuasion and significance to knowledge or decisions that result from big-data-based processes. These issues call for an active human role by creating opportunities to incorporate, in the most unbiased way, human expertise and prior knowledge in decision making and knowledge production. This also requires putting in place robust monitoring and appraisal mechanisms to ensure that relevant data is answering the right questions. As the proliferation of data continues to grow, we need to rethink the way we interact with data to serve human needs.

Particular textures of the minimal seesaw model

Following the Occam's razor principle, we explore the particular textures (featuring texture zeros and equalities) of the Dirac neutrino mass matrix in the minimal seesaw model (where only two right-handed neutrinos N1 and N2 are responsible for the generation of the light neutrino masses), in light of current neutrino experimental results and leptogenesis. The study will be performed for two particular patterns of the Majorana mass matrix MR for N1 and N2: (A) MR being diagonal (M100M2); (B) MR being of the form (0MM0).

Use of the Fundamental Problem of Statistics to Define the Validity Limit of Occam's Razor Principle

ABSTRACT: In statistics, to evaluate the significance of a result, one of the most used methods is the statistical hypothesis test. Using this theory, the fundamental problem of statistics can be expressed as follows: A statistical data does not represent useful information, but becomes useful information only when it is shown that it was not obtained Consequently, according to this point of view, among the hypotheses that perform the same prediction, we must choose the result that has a lower probability of being produced randomly. Therefore, the fundamental aspect of this approach is to calculate correctly this probability value. This problem is addressed by redefining what is meant by hypothesis. The traditional approach considers the hypothesis as the set of rules that actively participate in the forecast. Instead, we consider as hypotheses the sum of all the hypotheses made, also considering the hypotheses preceding the one used. Therefore, each time a prediction is made, our hypothesis increases in complexity and consequently increases its ability to adapt to a random data set. In this way, the complexity of a hypothesis can be precisely determined only if all previous attempts are known. Consequently, Occam's razor principle no longer has a general value, but its application depends on the information we have on the tested hypotheses.

The effects of air pollution sources / sensor array configurations on the likelihood of obtaining accurate source term estimations

Estimating the source term in the case of multiple leaks using a sparse sensor array is a challenging task. Here, the effect of sensor array/leak configurations on the reliability of the source term estimation is studied using two new measures. The first describes the overall change in the sensor array response to different source terms. The second represents the effect of the source term on the readout of each sensor in the array. These measures are subjected to several model cases differing in sensor array/leak configurations. Then, the source term is estimated using a self-adaptive multiobjective evolutionary (MOEA) search algorithm combined with a gas dispersion model. The method searches for a set of leaks, each one of which has a typical emission rate and location that results in a minimal difference between the sensors' actual and computed pollution concentration. This objective, which is often used for source term estimation, is traded off against the second objective of maintaining a minimum number of active sources, which follows Occam's razor principle of parsimony. Analysis of the results obtained for these model cases suggests that the measures can be implemented as a design tool using a combination of computer simulation and field experiments before operational deployment.

An innovative idea for low cost Braille e-reader

Background/Objectives: The aim of this study was to implement the rotating cubes for designing the Braillerefreshable display. The rotating cube Braille display has many benefits i.e., compact size, low cost etc. Methods/Statistical analysis: refreshable Braille display is an electro mechanical device which forms Braille charactersby raising pins through holes on flat surface. The raising of such pins is accomplished by complex and costly systems like solenoids or piezo-electric crystals or EAP (Electro Active Polimer). We developed the prototype of e-book reader, which is a low cost electro mechanical system aimed at assisting visually impaired to access the e-books. Findings: Many design concepts were proposed earlier and prototypes were built, and were tested successfully in thepast by using three simple rotating cubes. The main parameters like dot spacing, overall dimension, touching force and operating voltage obtained are discussed. Dot spacing 2.5 [mm] with reference to International standard has been maintained. Regarding overall dimension, the display unit is designed to have dimension of 120mm width to 150mm height adhering to NLS standard of Braille dot spacing, and the tactile surface will feature 13 lines with 14 characters per line. The touching force of 0.32N and operating voltage below 25 V was obtained and used. Novelty/Applications: The Solenoid is used for positioning the faces of individual cubes, which can be arranged in different positions thereby forming all Braille characters. The high power consumption which remained the major head ache in solenoid electromagnetic actuator models was a major challenge we faced initially, which we overcame with our unique innovative design combining cubes that can rotate to form Braille dots and electromagnetic actuator which can be positioned at various places to run this cubes. So Keeping Occam\'s Razor Principle in mind, we went ahead to design a multicell/multi-line electronic refreshable Braille display as a technology demonstrator by using three simple rotating cubes. Keywords: Braille cell; Rotating cubes; Actuators; Refreshable Braille displays

Parsimonious design of pattern recognition systems for slope stability analysis

State-of-the-art machine learning methods, such as (deep) neural networks and support vector classifiers, have been successfully applied to problems related to the prediction of natural hazard events. However, the effectiveness of those methods is, in general, obtained at the cost of a complex algorithmic architecture that requires fine-tuning of several parameters. Moreover, their current popularity as a hot topic in the literature makes researchers to ignore simpler and classical methods that might perform equally well and should be preferred according to the Occam’s razor principle. In this paper we exemplify this case by showing that two particular problems of slope stability prediction —that were recently solved using complex approaches named bee colony optimized support vector classifier and metaheuristic-optimized least squares support vector classifier, respectively— can be successfully solved by much more simpler pattern recognition methods. We also emphasize on the importance of data visualization and incremental evaluation during the design cycle of a parsimonious pattern recognition system.

Rule-based meta-analysis reveals the major role of PB2 in influencing influenza A virus virulence in mice

BackgroundInfluenza A virus (IAV) poses threats to human health and life. Many individual studies have been carried out in mice to uncover the viral factors responsible for the virulence of IAV infections. Nonetheless, a single study may not provide enough confident about virulence factors, hence combining several studies for a meta-analysis is desired to provide better views. For this, we documented more than 500 records of IAV infections in mice, whose viral proteins could be retrieved and the mouse lethal dose 50 or alternatively, weight loss and/or survival data, was/were available for virulence classification.ResultsIAV virulence models were learned from various datasets containing aligned IAV proteins and the corresponding two virulence classes (avirulent and virulent) or three virulence classes (low, intermediate and high virulence). Three proven rule-based learning approaches, i.e., OneR, JRip and PART, and additionally random forest were used for modelling. PART models achieved the best performance, with moderate average model accuracies ranged from 65.0 to 84.4% and from 54.0 to 66.6% for the two-class and three-class problems, respectively. PART models were comparable to or even better than random forest models and should be preferred based on the Occam’s razor principle. Interestingly, the average accuracy of the models was improved when host information was taken into account. For model interpretation, we observed that although many sites in HA were highly correlated with virulence, PART models based on sites in PB2 could compete against and were often better than PART models based on sites in HA. Moreover, PART had a high preference to include sites in PB2 when models were learned from datasets containing the concatenated alignments of all IAV proteins. Several sites with a known contribution to virulence were found as the top protein sites, and site pairs that may synergistically influence virulence were also uncovered.ConclusionModelling IAV virulence is a challenging problem. Rule-based models generated using viral proteins are useful for its advantage in interpretation, but only achieve moderate performance. Development of more advanced approaches that learn models from features extracted from both viral and host proteins shall be considered for future works.

Rescoring of docking poses under Occam's Razor: are there simpler solutions?

Ligand affinity prediction from docking simulations is usually performed by means of highly empirical and diverse protocols. These protocols often involve the re-scoring of poses generated by a force field (FF) based Hamiltonian to provide either estimated binding affinities-or alternatively, some empirical goodness score. Re-scoring is performed by so-called scoring functions-typically, a reweighted sum of FF terms augmented by additional terms (e.g., desolvation/entropic penalty, hydrophobicity, aromatic interactions etc.). Sometimes, the scoring function actually drives ligand positioning, but often it only operates on the best scoring poses ranked top by the initial ligand positioning tool. In either of these rather intricate scenarios, scoring functions are docking-specific models, and most require machine-learning-based calibration. Therefore, docking simulations are less straightforward when compared to "standard" molecular simulations in which the FF Hamiltonian defines the energy, and affinity emerges as an ensemble average property over pools of representative conformers (i.e., the trajectory). Paraphrasing on Occam's Razor principle, additional model complexity is only acceptable if demonstrated to bring a significant improvement of prediction quality. In this work we therefore examined whether the complexity inherent to scoring functions is indeed justified. For this purpose we compared sampler for multiple protein-ligand entities, a general purpose conformation sampler based on the AMBER/GAFF FF, complemented with continuum solvation terms, with several state of the art docking tools that rely on calibrated scoring functions (Glide, Gold, Autodock-Vina) in terms of its ability to top-rank the actives from large and diverse ligand series associated with various proteins. There is no clear winner of this study, where each program performed well on most of the targets, but also failed with respect to at least one of them. Therefore, a well-parameterized force field with a simple, energy-based ligand ranking protocol appears to be anas effective docking protocol as intricate rescoring strategies based on scoring functions. Atool that can sample the conformational space of the free ligand, the bound ligand and the protein binding site using the same force field may avoidmany of the approximations common to contemporary docking protocols and allow e.g., for docking into highly flexible active sites, when current scoring functions are not well suited to estimate receptor strain energies.

Maxwell’s Equations and Occam’s Razor

In this paper a straightforward application of Occam’s razor principle to Maxwell’s equation shows that only one entity, the electro- magnetic four-potential, is at the origin of a plurality of concepts and entities in physics. The application of the so called “Lorenz gauge” in Maxwell’s equations denies the status of real physical entity to a scalar field that has a gradient in space-time with clear physical meaning: the four-current density field. The mathematical formalism of space-time Clifford algebra is introduced and then used to encode Maxwell’s equations starting only from the electromagnetic four-potential. This approach suggests a particular Zitterbewegung (ZBW) model for charged elementary particles.

Fourier Spectroscopy: A Bayesian Way

The concepts of standard analysis techniques applied in the field of Fourier spectroscopy treat fundamental aspects insufficiently. For example, the spectra to be inferred are influenced by the noise contribution to the interferometric data, by nonprobed spatial domains which are linked to Fourier coefficients above a certain order, by the spectral limits which are in general not given by the Nyquist assumptions, and by additional parameters of the problem at hand like the zero-path difference. To consider these fundamentals, a probabilistic approach based on Bayes’ theorem is introduced which exploits multivariate normal distributions. For the example application, we model the spectra by the Gaussian process of a Brownian bridge stated by a prior covariance. The spectra themselves are represented by a number of parameters which map linearly to the data domain. The posterior for these linear parameters is analytically obtained, and the marginalisation over these parameters is trivial. This allows the straightforward investigation of the posterior for the involved nonlinear parameters, like the zero-path difference location and the spectral limits, and hyperparameters, like the scaling of the Gaussian process. With respect to the linear problem, this can be interpreted as an implementation of Ockham’s razor principle.

Trios\u2014promising in silico biomarkers for differentiating the effect of disease on the human microbiome network

Recent advances in the HMP (human microbiome project) research have revealed profound implications of the human microbiome to our health and diseases. We postulated that there should be distinctive features associated with healthy and/or diseased microbiome networks. Following Occam’s razor principle, we further hypothesized that triangle motifs or trios, arguably the simplest motif in a complex network of the human microbiome, should be sufficient to detect changes that occurred in the diseased microbiome. Here we test our hypothesis with six HMP datasets that cover five major human microbiome sites (gut, lung, oral, skin, and vaginal). The tests confirm our hypothesis and demonstrate that the trios involving the special nodes (e.g., most abundant OTU or MAO, and most dominant OTU or MDO, etc.) and interactions types (positive vs. negative) can be a powerful tool to differentiate between healthy and diseased microbiome samples. Our findings suggest that 12 kinds of trios (especially, dominantly inhibitive trio with mixed strategy, dominantly inhibitive trio with pure strategy, and fully facilitative strategy) may be utilized as in silico biomarkers for detecting disease-associated changes in the human microbiome, and may play an important role in personalized precision diagnosis of the human microbiome associated diseases.

Surfing on Protein Waves: Proteophoresis as a Mechanism for Bacterial Genome Partitioning.

Efficient bacterial chromosome segregation typically requires the coordinated action of a three-component machinery, fueled by adenosine triphosphate, called the partition complex. We present a phenomenological model accounting for the dynamic activity of this system that is also relevant for the physics of catalytic particles in active environments. The model is obtained by coupling simple linear reaction-diffusion equations with a proteophoresis, or "volumetric" chemophoresis, force field that arises from protein-protein interactions and provides a physically viable mechanism for complex translocation. This minimal description captures most known experimental observations: dynamic oscillations of complex components, complex separation, and subsequent symmetrical positioning. The predictions of our model are in phenomenological agreement with and provide substantial insight into recent experiments. From a nonlinear physics view point, this system explores the active separation of matter at micrometric scales with a dynamical instability between static positioning and traveling wave regimes triggered by the dynamical spontaneous breaking of rotational symmetry.

When Ockham razor's principle is not applicable: Differential diagnosis of a rare case of child and adolescent psychosis

IntroductionThe diagnosis of schizophrenia in children is rare. Less than 4% of schizophrenic patients begin before age 15 being much less stable than in adults as an entity in time. It is estimated that only 50% of diagnoses of schizophrenia in patients under 15 years are maintained over time. The most frequent differential diagnoses are bipolar disorder, post-traumatic stress disorder and dissociative disorder.Objective and methodsA case of a patient of 18 years old admitted in our service with diagnosis of paranoid schizophrenia due to the presence of delusional symptoms at age of 14 and due his evolution with impaired overall performance is presented. Upon arrival he presented delusions, self-referentiality and a strange phenotype with a pitched voice. Clinical history included presence of sexual abuse prior to debut of psychotic symptoms and rare medical comorbidity (diagnosed at age 15 of hypertension and paroxysmal sinus tachycardia). A kariotipe was done in a previous admission with normal results.ResultsDuring hospitalization symptomatic remission was achieved in just two days by decreasing antipsychotic potency of the treatment, he also presented elevated metanephrines and also elevated plasma aldosterone and renin in blood tests.ConclusionsWe discuss the differential diagnosis including schizophrenia, post-traumatic stress disorder with dissociative symptoms and endocrine pathology (pheochromocytoma and hyperaldosteronism).L. Galindo is a Rio Hortega fellowship (ISC III; CM14/00111).Disclosure of interestThe authors have not supplied their declaration of competing interest.

CNS Multiparameter Optimization Approach: Is it in Accordance with Occam's Razor Principle?

A detailed analysis of the possibility of using the Multiparameter Optimization approach (MPO) for CNS/non-CNS classification of drugs was carried out. This work has shown that MPO descriptors are able to describe only part of chemical transport in the CNS connected with transmembrane diffusion. Hence the "intuitive" CNS MPO approach with arbitrary selection of descriptors and calculations of score functions, search of thresholds of classification, and absence of any chemometric procedures, leads to rather modest accuracy of CNS/non-CNS classification models.

Evolution of Cuticular Hydrocarbons in the Hymenoptera: a Meta-Analysis.

Chemical communication is the oldest form of communication, spreading across all forms of life. In insects, cuticular hydrocarbons (CHC) function as chemical cues for the recognition of mates, species, and nest-mates in social insects. Although much is known about the function of individual hydrocarbons and their biosynthesis, a phylogenetic overview is lacking. Here, we review the CHC profiles of 241 species of Hymenoptera, one of the largest and most important insect orders, which includes the Symphyta (sawflies), the polyphyletic Parasitica (parasitoid wasps), and the Aculeata (wasps, bees, and ants). We investigated whether these taxonomic groups differed in the presence and absence of CHC classes and whether the sociality of a species (solitarily vs. social) had an effect on CHC profile complexity. We found that the main CHC classes (i.e., n-alkanes, alkenes, and methylalkanes) were all present early in the evolutionary history of the Hymenoptera, as evidenced by their presence in ancient Symphyta and primitive Parasitica wasps. Throughout all groups within the Hymenoptera, the more complex a CHC the fewer species that produce it, which may reflect the Occam’s razor principle that insects’ only biosynthesize the most simple compound that fulfil its needs. Surprisingly, there was no difference in the complexity of CHC profiles between social and solitary species, with some of the most complex CHC profiles belonging to the Parasitica. This profile complexity has been maintained in the ants, but some specialization in biosynthetic pathways has led to a simplification of profiles in the aculeate wasps and bees. The absence of CHC classes in some taxa or species may be due to gene silencing or down-regulation rather than gene loss, as demonstrated by sister species having highly divergent CHC profiles, and cannot be predicted by their phylogenetic history. The presence of highly complex CHC profiles prior to the vast radiation of the social Hymenoptera indicates a ‘spring-loaded’ system where the diversity of CHC needed for the complex communication systems of social insects were already present for natural selection to act upon, rather than having evolved independently. This diversity may have aided the multiple independent evolution of sociality within the Aculeata.Electronic supplementary materialThe online version of this article (doi:10.1007/s10886-015-0631-5) contains supplementary material, which is available to authorized users.

SafeRazor: Metastability-Robust Adaptive Clocking in Resilient Circuits

Razor-based circuits can run faster or at a lower voltage than those designed to work at the worst case corner. However, all known implementations are prone to failures due to the non-deterministic timing behavior introduced by metastability, even in the case where sufficient time is left for resolution. This paper analyzes the causes why Razor-based circuits fail and proposes a new scheme combining the Razor principle with stoppable clocks in a GALS setting. This scheme avoids any timing failure due to metastability and does not require any checkpointing or pipeline restarting logic, other than the usual auxiliary latch to store valid data. The experiments show how the Razor principle can be extended to any generic logic circuit, and not just to microprocessors with sophisticated pipeline flush/recovery mechanisms. In this way, the performance/power benefits of Razor can be adopted without the complex architectural changes required by the various Razor schemes in the literature.

Structure learning and the Occam's razor principle: a new view of human function acquisition.

We often encounter pairs of variables in the world whose mutual relationship can be described by a function. After training, human responses closely correspond to these functional relationships. Here we study how humans predict unobserved segments of a function that they have been trained on and we compare how human predictions differ to those made by various function-learning models in the literature. Participants' performance was best predicted by the polynomial functions that generated the observations. Further, participants were able to explicitly report the correct generating function in most cases upon a post-experiment survey. This suggests that humans can abstract functions. To understand how they do so, we modeled human learning using an hierarchical Bayesian framework organized at two levels of abstraction: function learning and parameter learning, and used it to understand the time course of participants' learning as we surreptitiously changed the generating function over time. This Bayesian model selection framework allowed us to analyze the time course of function learning and parameter learning in relative isolation. We found that participants acquired new functions as they changed and even when parameter learning was not completely accurate, the probability that the correct function was learned remained high. Most importantly, we found that humans selected the simplest-fitting function with the highest probability and that they acquired simpler functions faster than more complex ones. Both aspects of this behavior, extent and rate of selection, present evidence that human function learning obeys the Occam's razor principle.

Acute Coagulopathy of Trauma in the Rat. Shock 39

To the Editor Darlington and colleagues (1) at the US Army Institute of Surgical Research are to be congratulated in attempting to develop a rat model of acute coagulopathy resulting from polytrauma and hemorrhage. As the authors point out, coagulopathy occurs early in hemorrhagic trauma and is a major contributor to mortality and morbidity. Coagulopathic bleeding was first reported in the Vietnam War (2) and rediscovered some 40 years later by Brohi (3). Several large trials have shown that 24% to 36% of severely injured patients have acute traumatic coagulopathy on admission to the emergency department (mean arrival time, 75 min), and it was associated with a 4-fold increase in mortality (3–5). In 2011, and 2012, we reported profound hypocoagulability in nonheparinized, mild accidental hypothermic (~34°C), male Sprague-Dawley rats during 20-min pressure-controlled bleeding (∼40% blood loss) and 60-min shock (6, 7). We further showed that 0.3 mL intravenous bolus of 7.5% NaCl adenosine-lidocaine/Mg2+ (ALM) raised mean arterial pressure into the hypotensive range, and it fully corrected PT and aPTT to baseline levels at 60-min resuscitation (7). Darlington and colleagues also showed increases in PT and aPTT in their rat model, although these were markedly reduced in comparison to our study. With the exception of minor surgery for ventilation tube insertion, the 10-fold increases in PT and aPTT we argued in our study were related to (1) 40% blood loss and (2) the emergent shock state and tissue hypoperfusion (7). A surprising result from the Darlington study was that while increases in PT and aPTT indicate hypocoagulopathy (Fig. 2), they report rotational thromboelastometry (ROTEM) shortening of ExTEM clotting time (CT) indicating hypercoagulability. Although there was a dip in aPTT at 120 min (Fig. 2), this fall was a return to baseline, not below baseline to support a hypercoagulable state. Was this shortening in ExTEM CT found in InTEM CT, because aPTT is equivalent to InTEM, not ExTEM? Were there differences in InTEM and ExTEM mean clot firmness? What was the core temperature in their rat model of coagulopathy over the 300 min, as this was not specified, and changes in body temperature have a profound effect on coagulation. In 2013, Park and colleagues (8) demonstrated in rats exposed to uncontrolled blood loss that their blood clotted less firmly during traumatic hemorrhage, and hypothermia prolonged CTs. Another concern with the study of Darlington and colleagues is the complexity of their results, which they acknowledge, and this may be reflected in the complexity of their model, which involved hemorrhage after crush injuries to the small intestine and right and left liver lobes, right leg skeletal muscle, and a right femur fracture. Animal models to test novel resuscitation strategies must be clinically and battlefield relevant (9); however, perhaps we should all take heed of Ockham’s razor principle of economy, “Plurality should not be posited without necessity,” when attempting to unravel the underlying mechanisms of acute coagulopathy of trauma. Hayley L. Letson Geoffrey P. Dobson Heart and Trauma Research Laboratory Department of Physiology and Pharmacology School of Biomedical Sciences James Cook University Townsville, Queensland Australia

Numerical models based on a minimal set of sarcolemmal electrogenic proteins and an intracellular Ca2 + clock generate robust, flexible, and energy-efficient cardiac pacemaking

Recent evidence supports the idea that robust and, importantly, FLEXIBLE automaticity of cardiac pacemaker cells is conferred by a coupled system of membrane ion currents (an “M-clock”) and a sarcoplasmic reticulum (SR)-based Ca2+ oscillator (“Ca2+clock”) that generates spontaneous diastolic Ca2+ releases. This study identified numerical models of a human biological pacemaker that features robust and flexible automaticity generated by a minimal set of electrogenic proteins and a Ca2+clock. Following the Occam's razor principle (principle of parsimony), M-clock components of unknown molecular origin were excluded from Maltsev–Lakatta pacemaker cell model and thirteen different model types of only 4 or 5 components were derived and explored by a parametric sensitivity analysis. The extended ranges of SR Ca2+ pumping (i.e. Ca2+clock performance) and conductance of ion currents were sampled, yielding a large variety of parameter combination, i.e. specific model sets. We tested each set's ability to simulate autonomic modulation of human heart rate (minimum rate of 50 to 70bpm; maximum rate of 140 to 210bpm) in response to stimulation of cholinergic and β-adrenergic receptors. We found that only those models that include a Ca2+clock (including the minimal 4-parameter model “ICaL+IKr+INCX+Ca2+clock”) were able to reproduce the full range of autonomic modulation. Inclusion of If or ICaT decreased the flexibility, but increased the robustness of the models (a relatively larger number of sets did not fail during testing). The new models comprised of components with clear molecular identity (i.e. lacking IbNa & Ist) portray a more realistic pacemaking: A smaller Na+ influx is expected to demand less energy for Na+ extrusion. The new large database of the reduced coupled-clock numerical models may serve as a useful tool for the design of biological pacemakers. It will also provide a conceptual basis for a general theory of robust, flexible, and energy-efficient pacemaking based on realistic components.

Crude oil price forecasting: Experimental evidence from wavelet decomposition and neural network modeling

Oil price prediction has usually proved to be an intractable task due to the intrinsic complexity of oil market mechanism. In addition, the recent oil shock and its consequences relaunch the debate on understanding the behavior underlying the expected oil prices. Combining the dynamic properties of multilayer back propagation neural network and the recent Harr A trous wavelet decomposition, a Hybrid model HTW-MPNN is implemented to achieve prominent prediction of crude oil price. While recent studies focus on the determination of the best forecasting model by comparing various neural architectures or applying several decomposition techniques to the ANN, the new insight of this paper is to target the issue of the transfer function selection providing robust simulations on both in sample and out of sample basis. Based on the work of Yonaba, H., Anctil, F., and Fortin, V. (2010) “Comparing Sigmoid Transfer Functions for Neural Network Multistep Ahead Stream flow forecasting”. Journal of Hydrologic Engineering, April, 275–283, we use three variants of activation function namely sigmoid, bipolar sigmoid and hyperbolic tangent in order to test the model's flexibility. Furthermore, the forecasting robustness is checked through several levels of input–hidden nodes. Comparatively, results of HTW-MBPNN perform better than the conventional BPNN. Our conclusions add a major attribute to the previous studies corroborating the Occam razor's principle, especially when simulations are constructed through training and testing phases simultaneously. Finally, more eligible forecasting power is found according to the wavelet oil price signal which appears to be the closest to the real anticipations of future oil price fluctuations.