Non-uniform Data Distributions Research Articles

Gravity Predictions in Data-Missing Areas Using Machine Learning Methods

Gravity data, comprising a key foundational dataset, are crucial for various research, including land subsidence monitoring, geological exploration, and navigational positioning. However, the collection of gravity data in specific regions is difficult because of environmental, technical, and economic constraints, resulting in a non-uniform distribution of the observational data. Traditionally, interpolation methods such as Kriging have been widely used to deal with data gaps; however, their predictive accuracy in regions with sparse data still needs improvement. In recent years, the rapid development of artificial intelligence has opened up a new opportunity for data prediction. In this study, utilizing the EGM2008 satellite gravity model, we conducted a comprehensive analysis of three machine learning algorithms—random forest, support vector machine, and recurrent neural network—and compared their performances against the traditional Kriging interpolation method. The results indicate that machine learning methods exhibit a marked advantage in gravity data prediction, significantly enhancing the predictive accuracy.

Enhancing Bucket Sort Efficiency: A Combinatorial Approach to Algorithm Optimization

Abstract: The efficiency of sorting algorithms is critical to numerous computational processes, including data management, search operations, and real-time applications. Among these algorithms, Bucket Sort is notable for its linear-time performance when data is uniformly distributed over a defined range. However, its efficiency suffers significantly with non-uniform data distributions, resulting in unbalanced buckets and increased sorting time. In our research, we explore how combinatorial algorithms can enhance Bucket Sort’s efficiency by optimizing bucket partitioning, dynamic allocation, and internal sorting strategies. Through rigorous mathematical proofs and analysis, we demonstrate that the application of combinatorial techniques yields substantial improvements in average-case performance, particularly in handling skewed data distributions. Practical case studies in large-scale data processing illustrate the adaptability and effectiveness of these optimizations, supporting their potential in real-world applications.

Empowering precise advertising with Fed-GANCC: A novel federated learning approach leveraging Generative Adversarial Networks and group clustering.

In the realm of targeted advertising, the demand for precision is paramount, and the traditional centralized machine learning paradigm fails to address this necessity effectively. Two critical challenges persist in the current advertising ecosystem: the data privacy concerns leading to isolated data islands and the complexity in handling non-Independent and Identically Distributed (non-IID) data and concept drift due to the specificity and diversity in user behavior data. Current federated learning frameworks struggle to overcome these hurdles satisfactorily. This paper introduces Fed-GANCC, an innovative federated learning framework that synergizes Generative Adversarial Networks (GANs) and Group Clustering. The framework incorporates a user data augmentation algorithm predicated on adversarial generative networks to enrich user behavior data, curtail the impact of non-uniform data distribution, and enhance the applicability of the global machine learning model. Unlike traditional approaches, our framework offers user data augmentation algorithms based on adversarial generative networks, which not only enriches user behavior data but also reduces the challenges posed by non-uniform data distribution, thereby enhancing the applicability of the global machine learning (ML) model. The effectiveness of Fed-GANCC is distinctly showcased through experimental results, outperforming contemporary methods like FED-AVG and FED-SGD in terms of accuracy, loss value, and receiver operating characteristic (ROC) indicators within the same computing time. Experimental results vindicate the effectiveness of Fed-GANCC, revealing substantial enhancements in accuracy, loss value, and receiver operating characteristic (ROC) metrics compared to FED-AVG and FED-SGD given the same computational time. These outcomes underline Fed-GANCC's exceptional prowess in mitigating issues such as isolated data islands, non-IID data, and concept drift. With its novel approach to addressing the prevailing challenges in targeted advertising such as isolated data islands, non-IID data, and concept drift, the Fed-GANCC framework stands as a benchmark, paving the way for future advancements in federated learning solutions tailored for the advertising domain. The Fed-GANCC framework promises to offer pivotal insights for the future development of efficient and advanced federated learning solutions for targeted advertising.

An algorithm for constructing spatial vector data storage based on KD-tree and density estimation

With the widespread application of spatial vector data in various fields, this paper proposes a novel algorithm for constructing spatial vector data storage. The algorithm is based on KD-tree and density estimation techniques, aiming to improve the storage efficiency and query performance of large-scale spatial vector data. The algorithm first efficiently organizes spatial vector data using KD-tree, and then performs density estimation based on the Locality Sensitive Hashing (LSH) algorithm. Algorithm optimizations for spatial partitioning are applied to the KD-tree construction algorithm, reducing the search range during queries and improving retrieval speed. By testing with Green Tide data, the algorithm based on KD-tree and density estimation demonstrates higher query efficiency and better scalability across multiple test datasets. Particularly, when dealing with large-scale datasets characterized by non-uniform data distributions, this algorithm significantly improves data retrieval speed while maintaining low storage overhead.

Video-based beat-by-beat blood pressure monitoring via transfer deep-learning

Currently, learning physiological vital signs such as blood pressure (BP), hemoglobin levels, and oxygen saturation, from Photoplethysmography (PPG) signal, is receiving more attention. Despite successive progress that has been made so far, continuously revealing new aspects characterizes that field as a rich research topic. It includes a diverse number of critical points represented in signal denoising, data cleaning, employed features, feature format, feature selection, feature domain, model structure, problem formulation (regression or classification), and model combinations. It is worth noting that extensive research efforts are devoted to utilizing different variants of machine learning and deep learning models while transfer learning is not fully explored yet. So, in this paper, we are introducing a per-beat rPPG-to-BP mapping scheme based on transfer learning. An interesting representation of a 1-D PPG signal as a 2-D image is proposed for enabling powerful off-the-shelf image-based models through transfer learning. It resolves limitations about training data size due to strict data cleaning. Also, it enhances model generalization by exploiting underlying excellent feature extraction. Moreover, non-uniform data distribution (data skewness) is partially resolved by introducing logarithmic transformation. Furthermore, double cleaning is applied for training contact PPG data and testing rPPG beats as well. The quality of the segmented beats is tested by checking some of the related quality metrics. Hence, the prediction reliability is enhanced by excluding deformed beats. Varying rPPG quality is relaxed by selecting beats during intervals of the highest signal strength. Based on the experimental results, the proposed system outperforms the state-of-the-art systems in the sense of mean absolute error (MAE) and standard deviation (STD). STD for the test data is decreased to 5.4782 and 3.8539 for SBP and DBP, respectively. Also, MAE decreased to 2.3453 and 1.6854 for SBP and DBP, respectively. Moreover, the results for BP estimation from real video reveal that the STD reaches 8.027882 and 6.013052 for SBP and DBP, respectively. Also, MAE for the estimated BP from real videos reaches 7.052803 and 5.616028 for SBP and DBP, respectively.Graphical abstractProposed camera-based blood pressure monitoring system

Spatial-temporal traffic data imputation based on dynamic multi-level generative adversarial networks for urban governance

Complete and accurate traffic data lays an essential foundation for urban governance and intelligent transportation system. Data missing is an inevitable issue in road network analysis with the expansion of urban roads and increasing application of data monitors. Data imputation is regarded as a cost-effective approach for solving data missing problem. However, the spatial-temporal features of massive traffic data have not been comprehensively considered in the previous related studies, and non-uniform distribution of missing data often causes poor results. Therefore, we propose a dynamic multi-level generative adversarial networks (MLGAN) model for the imputation of large-scale spatial-temporal traffic data. The model first captures the spatial-temporal features of traffic data through composite features extractor based on the bidirectional gated recurrent unit (BGRU) and the graph convolutional network (GCN) structure. Then, different sampler structures with varied missing rate are adopted to analyze the feature data and present the reconstruction results. Enhanced by the dynamic stratification strategy, the model is less sensitive to the variation of missing rates. Finally, data imputation experiments are conducted to validate the proposed approach using the data collected from the PeMS dataset. The results show that the proposed model has superior data imputation performance on large-scale dataset compared with the traditional models.

Real-time phase-resolved ocean wave prediction in directional wave fields: Enhanced algorithm and experimental validation

This paper details a comprehensive study of deterministic real-time wave forecasting in directional seas. By using wave models on the basis of a Lagrangian description, a good balance was achieved between computational efficiency and model accuracy. Nevertheless, due to the highly non-uniform spatial distribution of data and the relatively small size of data in time inherent to remote optical measurements, the initial conditions are determined through an optimization process, which is computationally demanding, especially in multidirectional sea states. Accordingly, in order to offer a real-time system of wave prediction in the case of multidirectional waves, we propose a simplified and succinct assimilation method for the process of wave reconstruction. We also develop a three-dimensional spatio-temporal prediction zone where the future evolution of wave fields can be estimated based on wave measurements. Lastly, we outline a tank-scale experimental campaign conducted to mimic the measurements of a LIDAR in a real configuration. A comparison of model performances with the experimental observations shows that in a multidirectional approach, it is necessary to consider wave components in direction as well as in frequency to achieve nearly the same accuracy as for unidirectional seas.

A Query Optimizer for Range Queries over Multi-Attribute Trajectories

A multi-attribute trajectory consists of a spatio-temporal trajectory and a set of descriptive attributes. Such data enrich the representation of traditional spatio-temporal trajectories to have comprehensive knowledge of moving objects. Range query is a fundamental operator over multi-attribute trajectories. Such a query contains two predicates, spatio-temporal and attribute, and returns the objects whose locations are within a distance threshold to the query trajectory and attributes contain expected values. There are different execution plans for answering the query. To enhance the capability of a trajectory database, an optimizer is essentially required to (i) accurately estimate the cost for alternative query strategies in terms of disk accesses, (ii) build a decision-making module that automatically sorts the data in an appropriate way and selects the optimal query plan, and (iii) update the analytical models when new trajectories are arrived. The cost model supports both uniform and non-uniform spatio-temporal data distribution and incorporates attribute distribution. The optimizer is fully developed inside a database system kernel and comprehensively evaluated in terms of accuracy and effectiveness by using large real and synthetic datasets.

Evaluation Platform for DDM Algorithms With the Usage of Non-Uniform Data Distribution Strategies

Huge amounts of data are collected in numerous independent data storage facilities around the world. However, how the data is distributed between physical locations remains unspecified. Downloading all of the data for the purpose of processing it is undesirable and sometimes even impossible. Various methods have been proposed for performing data mining tasks, but the main problem is the lack of an objective strategy for comparing them. The authors present current research on a novel evaluation platform for distributed data mining (DDM) algorithms. The proposed platform opens up a new field to evaluate algorithms in terms of the quality of the results, transfer used, and speed, but also for the use of a non-uniform data distribution among independent nodes during algorithm evaluation. This work introduces a ‘data partitioning strategy’ term referring to a specific, not necessarily uniform data distribution. A brief evaluation for three clustering algorithms is also reported, showing the usability and simplicity of identifying differences in processing with the use of the platform.

High-Throughput Parallel SRAM-Based Hash Join Architecture on FPGA

The hash join operator is one of the most important relational operations used in database. The offloading and acceleration of this operation on hardware has been a technique of growing interest for a long time. However, the non-uniform distribution of data caused by hash collisions negatively affects the throughput of the hash join algorithm, owing the variation in the number of hash table accesses required for each lookup. To resolve this issue, a non-collision parallel static random-access memory (SRAM)-based hash join architecture is presented. This architecture utilizes multiple hash functions and content addressable memories (CAMs) to eliminate hash collision, thereby ensuring a worst constant memory access for each phase in the hash join algorithm and consequently improving the hash join throughput. The proposed architecture was implemented on a Xilinx field programmable gate array (FPGA), and the experimental results show that our design achieved a high hash join throughput of 153.6 million tuples per second, and a speedup factor of at least 2.5 with the best existing FPGA-based hash join architecture and a match rate of 50%.

Data-parallel query processing on non-uniform data

Graphics processing units (GPUs) promise spectacular performance advantages when used as database coprocessors. Their massive compute capacity, however, is often hampered by control flow divergence caused by non-uniform data distributions. When data-parallel work items demand for different amounts or types of processing, instructions execute with lowered efficiency. Query compilation techniques---a recent advance in GPU-accelerated database processing---suffer from the problem even more, because divergence effects are amplified during the execution of fused pipeline operators. In this work, we identify two types of control flow divergence--- filter divergence and expansion divergence ---that frequently occur in real world workloads. We quantify the problem for two poster cases and propose techniques to balance these divergence effects. By balancing divergence effects, our approach is able to restore processing efficiency even when pipelines contain heavily skewed operations. Our query compiler DogQC has a wider range of functionality than other query coprocessors and achieves performance improvements. We observe shorter execution times for TPC-H benchmark queries by factors up to 4.51x compared with existing GPU query compilers and by factors up to 4.54x compared with CPU-based systems.

Numerical representation of rainfall field in basins of the Upper Jordan River and of the Yarmouk River

A non-linear complex rain–elevation regression model is proposed for the numerical representation of rainfall field in an area with complex plain/mountainous topography, non-uniform distribution of rain gauge network and scarce data of observations. The model is applied to basins of the Upper Jordan River (of the Lake Kinneret) and of the Yarmouk River in which the major flow of the Jordan River is formed. The model is implemented in two steps: (1) study of the rainfall field, based on the average long-term (climatic) data, its description by the function of elevation and other factors, and optimization of model parameters (normalized coefficients of the Taylor series); and (2) estimation of rainfall in each historical year using the available data (less complete and irregular than climatic data) as well as a priori known parameters. The basic hypothesis is inter-annual stability of the model parameters. As a necessary primary stage, the Upper Jordan River (Lake Kinneret) Basin and the Yarmouk River Basin were divided, respectively, into seven and five regions, considering the specific regional relationships between the mean annual rain depth and elevation. It occurred that both basins are represented by a common system of rain–elevation curves as a single rain field where the mean annual rain increases with altitude and decreases from west to east and from north to south. For each region, the parameters of the model were optimized as a base for estimating the annual (month) rain (mean or yearly volume/depth) in each point of grid, in each basin (sub-basin) or in the whole watershed. The necessary condition is numerical presentation of the topography. Derived rain rates can serve as fundamental input data for numerical modeling of surface- and groundwater flow. This method can be applied to other areas at different temporal and spatial scales.

Possible Impact of Spatial and Temporal Non-Uniformity in Land Surface Temperature Data on Trend Estimation

The present work investigates possible impact of the non-uniformity in observed land surface temperature on trend estimation, based on Climatic Research Unit (CRU) Temperature Version 4 (CRUTEM4) monthly temperature datasets from 1900 to 2012. The CRU land temperature data exhibit remarkable non-uniformity in spatial and temporal features. The data are characterized by an uneven spatial distribution of missing records and station density, and display a significant increase of available sites around 1950. Considering the impact of missing data, the trends seem to be more stable and reliable when estimated based on data with < 40% missing percent, compared to the data with above 40% missing percent. Mean absolute error (MAE) between data with < 40% missing percent and global data is only 0.011°C (0.014°C) for 1900–50 (1951–2012). The associated trend estimated by reliable data is 0.087°C decade–1 (0.186°C decade–1) for 1900–50 (1951–2012), almost the same as the trend of the global data. However, due to non-uniform spatial distribution of missing data, the global signal seems mainly coming from the regions with good data coverage, especially for the period 1900–50. This is also confirmed by an extreme test conducted with the records in the United States and Africa. In addition, the influences of spatial and temporal non-uniform features in observation data on trend estimation are significant for the areas with poor data coverage, such as Africa, while insignificant for the countries with good data coverage, such as the United States.

Job Performance Optimization Method Based on Data Balance in the Wireless Sensor Networks

In the wireless sensor network, the representative MapReduce computing model based on data center has been widely used in large-scale data processing. In the data transmission phase, the wireless sensor network system uses the hash method to distribute data for each Reduce task based on the number of Reduce tasks. This data partitioning method based on the hash function results in non-uniform distribution of the output data in the data transmission phase and further leads to skewing of the input data in the Reduce task. Data skew will result in load imbalance in the Reduce phase and causes the system performance to degrade. In order to eliminate the data skew problem in the Reduce phase, this paper presents a load balancing method, which consists of two parts: the virtual partitioning method based on the consistent hashing and the heterogeneity-aware loads balancing (HLB) algorithm. The experimental results show that the proposed method can eliminate the data skew in the Reduce phase and distribute the load equitably for each Reduce task. In addition, the method produces less system overhead.

Searching for a Cosmological Preferred Direction with 147 Rotationally Supported Galaxies

Abstract It is well known that the Milgrom’s modified Newtonian dynamics (MOND) explains well the mass discrepancy problem in galaxy rotation curves. The MOND predicts a universal acceleration scale below which the Newtonian dynamics is still invalid. We get the universal acceleration scale of 1.02 × 10−10 m s−2 by using the Spitzer Photometry and Accurate Rotation Curves (SPARC) data set. Milgrom suggested that the acceleration scale may be a fingerprint of cosmology on local dynamics and related to the Hubble constant g † ∼ cH 0. In this paper, we use the hemisphere comparison method with the SPARC data set to investigate possible spatial anisotropy on the acceleration scale. It is found that the hemisphere of the maximum acceleration scale is in the direction , with g †,max = 1.10 × 10−10 m s−2, while the hemisphere of the minimum acceleration scale is in the opposite direction , with g †,min = 0.76 × 10−10 m s−2. The level of anisotropy reaches up to 0.37 ± 0.04. Robust tests show that such an anisotropy cannot be reproduced by a statistically isotropic data set. We also show that the spatial anisotropy on the acceleration scale is less correlated with the non-uniform distribution of the SPARC data points in the sky. In addition, we confirm that the anisotropy of the acceleration scale does not depend significantly on other physical parameters of the SPARC galaxies. It is interesting to note that the maximum anisotropy direction found in this paper is close with other cosmological preferred directions, particularly the direction of the “Australia dipole” for the fine structure constant.

Learning a Robust Local Manifold Representation for Hyperspectral Dimensionality Reduction

Local manifold learning has been successfully applied to hyperspectral dimensionality reduction in order to embed nonlinear and nonconvex manifolds in the data. Local manifold learning is mainly characterized by affinity matrix construction, which is composed of two steps: neighbor selection and computation of affinity weights. There is a challenge in each step: First, the neighbor selection is sensitive to complex spectral variability due to nonuniform data distribution, illumination variations, and sensor noise; second, the computation of affinity weights is challenging due to highly correlated spectral signatures in the neighborhood. To address the two issues, in this paper, a novel manifold learning methodology based on locally linear embedding is proposed through learning a robust local manifold representation. More specifically, a hierarchical neighbor selection is designed to progressively eliminate the effects of complex spectral variability using joint normalization and to robustly compute affinity (or reconstruction) weights reducing multicollinearity via the refined neighbor selection. Additionally, an idea that combines spatial–spectral information is introduced into the proposed manifold learning methodology to further improve the robustness of affinity calculations. Classification is explored as a potential application for validating the proposed algorithm. The classification accuracy in the use of different dimensionality reduction methods is evaluated and compared, while two kinds of strategies are applied in selecting the training and test samples: random sampling and region-based sampling. Experimental results show the classification accuracy obtained by the proposed method is superior to those state-of-the-art dimensionality reduction methods.

Exploiting coarse-grained reused-based opportunities in Big Data multi-query optimization

Multi-query optimization in Big Data becomes a promising research direction due to the popularity of massive data analytical systems (e.g., MapReduce and Flink). The multi-query is translated into jobs. These jobs are routinely submitted with similar tasks to the underling Big Data analytical systems. These similar tasks are considered complicated and computation overhead. Therefore, there are some existing techniques that have been proposed for exploiting sharing tasks in Big Data multi-query optimization (e.g., MRShare and Relaxed MRShare). These techniques are heavily tailored relaxed optimizing factors of fine-grained reused-based opportunities. In accordance with Big Data multi-query optimization, the existing fine-grained techniques are only concerned with equal tuples size and uniform data distribution. These issues are not applicable to the real-world distributed applications which depend on coarse-grained reused-based opportunities, such as non-equal tuples size and non-uniform data distribution. These two issues receive more-attention in Big Data multi-query optimization, to minimize the data read from or written back to Big Data infrastructures (e.g., Hadoop). In this paper, Multi-Query Optimization using Tuple Size and Histogram (MOTH) system has been proposed to consider the granularity of the reused-based opportunities. The proposed MOTH system exploits the coarse-grained of the fully and partially reused-based opportunities among queries with considering non-equal tuples size and non-uniform data distribution to avoid repeated computations. According to the proposed MOTH system, a combined technique has been introduced for estimating the coarse-grained reused-based opportunities horizontally and vertically. The horizontal estimation of non-equal tuples size has been done by extracting metadata in column-level, while the vertical estimation of non-uniform data distribution is concerned with using pre-computed histogram in row-level. In addition, the MOTH system estimates the coarse-grained reused-based opportunities with considering slow storage (i.e., limited physical resources or fewer allocated virtualized resources) to produce the accurate estimation of the reused results costs. Then, a cost-based heuristic algorithm has been introduced to select the best reused-based opportunity and generate an efficient multi-query execution plan. Because the partial reused-based opportunities have been considered, extra computations are needed to retrieve the non-derived results. Also, a partial reused-based optimizer has been tailored and added to the proposed MOTH system to reformulate the generated multi-query plan to improve the shared partial queries. According to the experimental results of the proposed MOTH system using TPC-H benchmark, it is found that multi-query execution time has been reduced by considering the granularity of the reused results.

A Variational Bayesian inference method for parametric imaging of PET data

In dynamic Positron Emission Tomography (PET) studies, compartmental models provide the richest information on the tracer kinetics of the tissue. Inverting such models at the voxel level is however quite challenging due to the low signal-to-noise ratio of the time activity curves. In this study, we propose the use of a Variational Bayesian (VB) approach to efficiently solve this issue and thus obtain robust quantitative parametric maps.VB was adapted to the non-uniform noise distribution of PET data. Moreover, we propose a novel hierarchical scheme to define the model parameter priors directly from the images in case such information are not available from the literature, as often happens with new PET tracers.VB was initially tested on synthetic data generated using compartmental models of increasing complexity, providing accurate (%bias<2%±2%, root mean square error<15%±5%) parameter estimates. When applied to real data on a paradigmatic set of PET tracers (L-[1-11C]leucine, [11C]WAY100635 and [18F]FDG), VB was able to generate reliable parametric maps even in presence of high noise in the data (unreliable estimates<11%±5%).

Interpolation in the eXtended Classifier System: An architectural perspective

Sparseness in the input space can be challenging for Learning Classifier Systems.Interpolation is proposed as a solution to tackle this problem.Two approaches to incorporate interpolation in the eXtended Classifier System are introduced.Experimental results promise beneficial effects during the learning process.The presented results improve the observations reported elsewhere. Machine Learning techniques constitute a key factor to make Organic Computing (OC) systems self-adaptive and self-reconfigurable at runtime. OC systems are therefore equipped with a so-called self-learning property enabling them to react appropriately when the environmental demands change and the system is faced possibly unforeseen situations. The eXtended Classifier System (XCS) is a rule-based evolutionary online learning system that has gained plenty attention in the research field of Genetic-based Machine Learning in general and within the OC initiative in particular. In this article, the XCS system is structurally extended to incorporate numerical interpolation. With the presented approaches we pursue the overall goal to overcome the challenge of sparsely distributed samples in the problem space resulting from e.g., non-uniform data distributions. A novel Interpolation Component (IC) is introduced and two architectural integration approaches are discussed. We elaborate on three strategies to integrate interpolated values into various algorithmic steps of XCS. The potential of incorporating interpolation techniques is underpinned by an evaluation on a rather challenging theoretical classification task, called the checkerboard problem.

Variations of the ionospheric parameters and vertical electron density distribution at the northern edge of the EIA from 2010 to 2015 along 95°E and comparison with the IRI-2012

The vertical electron density profiles over Dibrugarh (27.5°N, 95°E, 43° dip) a low mid latitude station normally located at the northern edge of the EIA for the period of July 2010 till October 2015 are constructed from the measured bottom side profiles and ionosonde-GPS TEC assisted Topside Sounder Model (TSM) topside profiles. The bottom side density profiles are obtained by using POLAN on the manually scaled ionograms. The topside is constructed by the modified ionosonde assisted TSM model (TaP-TSM assisted by POLAN) which is integrated with POLAN for the first time. The reconstructed vertical profile is compared with the IRI predicted density profile and the electron density profile obtained from the COSMIC/FORMOSAT radio occultation measurements over Dibrugarh. The bottom side density profiles are fitted to the IRI bottom side function to obtain best-fit bottom side thickness parameter B0 and shape parameter B1. The temporal and solar activity variation of the B-parameters over Dibrugarh are investigated and compared to those predicted by IRI-2012 model with ABT-2009 option. The bottom side thickness parameter B0 predicted by the IRI model is found to be similar to the B0 measured over Dibrugarh in the night time and the forenoon hours. Differences are observed in the early morning and the afternoon period. The IRI doesn’t reproduce the morning collapse of B0 and overestimates the B0 over Dibrugarh in the afternoon period, particularly in summer and equinox. The IRI model predictions are closest to the measured B0 in the winter of low solar activity. The B0 over Dibrugarh is found to increase by about 15% with solar activity during the period of study encompassing almost the first half of solar cycle 24 but solar activity effect was not observed in the B1 parameter. The topside profile obtained from TaP profiler is thicker than the IRI topside in equinox from afternoon to sunrise period but is similar to the IRI in summer daytime. The differences in the bottom side may be attributed to the non-inclusion of ground measurements from 90°E to 100°E longitude in the ABT-2009 model while differences in the topside could be due the non-uniform longitudinal distribution of topside sounder profiles data and the stronger fountain effect in this longitude.