Raw Code Research Articles

Tight integration of real-time GNSS PPP and INS: Ionosphere-free combined vs uncombined models

The GNSS Precise Point Positioning (PPP) model is usually established in either an ionospheric-free (IF) combined form or an uncombined (UC) form. These formulations can be equivalent in theory but their applications in practice could still perform differently when integrated with external sensors. In this study, we compared the positioning performance of the two PPP models tightly coupled with the Inertial Navigation System (INS) using a high-grade inertial measurement unit (IMU) in real vehicle navigation tests. The ambiguity resolution (AR) was also exploited in the two PPP models after applying the observable specific biases (OSB) to the GNSS raw code and phase measurements. According to the results, under good satellite observability the UC PPP/INS tightly-coupled integration (TCI) significantly outperforms the IF PPP/INS TCI. The UC TCI model with AR could achieve a positioning accuracy of 4.6 and 3.0 cm in the horizontal and vertical directions, which are improved by 37 % and 63 % respectively relative to the IF TCI model. However, in the case of frequent GNSS signal interruptions or poor satellite observation condition, the IF TCI model shows a superior reliability than the UC TCI. Nevertheless, when the ionospheric parameters are properly constrained in the UC TCI model, substantial improvements in terms of convergence and accuracy are obtained. The UC PPP augmented with external precise ionospheric information would greatly increase the cost, and users may select the appropriate PPP model with INS TCI in real applications in accordance with the demanded accuracy level and measuring conditions.

Impaired health in working children: a critical ethnography

Child labor is one of the important social issues that deprive children of many fundamental rights, and make them face many problems and consequences, including health problems. Thus, this study was conducted with the aim of examining the health of working children in Tehran. This is an ethnographic study that was conducted using Carspecken’s approach and was completed in 2022. The main participants of this study included working children aged 10–18 years living in Tehran. In order to collect information, the researcher was present at the workplace, school, and living places of working children for more than two years, observing their lives and activities. Formal and informal interviews were also conducted with the working children and informed people. In total, hundreds of working children were assessed and observed in this research. A friendly conversation was formed between the researcher and more than 50 children, and official interviews were conducted with six of the working children. Also, more than 10 official interviews were conducted with informed people and parents of working children. In addition to observations and interviews, documents such as medical records and drawings of working children were also examined and interpreted. The information obtained from observations, interviews, and documents was entered into MAXQDA software, and its raw codes were extracted. The high-level codes as well as sub and main categories were formed from the aggregation of low-level codes. Impaired health was formed from three subcategories of tormented body (work and environmental trauma, sexual abuse, malnutrition, fatigue, sleep disorder and inadequate hygiene), disquieted mind (anxious children, depression and isolation, reduced self-esteem and unfocused mind) and disrupted sociability (negative social role modeling, aggression and violence, stubbornness and vindictiveness, harassment and nuisance, reprehensible social behaviors, neglecting others’ ownership, disturbed relationships and out-group self-censorship). The results of the present study showed that the health of working children is compromised in various physical, psychological, and social ways. Therefore, some measures should be taken at the national and international levels to improve their health, such as revising the existing laws regarding children and informing children of their rights.

Software Defect Prediction and Automatic Repair Based on Machine Learning

Software defects can have a significant impact on quality, reliability, and development costs if not addressed properly. With the increasing complexity of modern software, traditional testing methods struggle to keep up. Machine learning and deep learning have emerged as promising new approaches for software defect prediction (SDP) and automatic repair. These approaches leverage historical data to identify patterns associated with defects. Deep learning models can automatically learn complex feature representations from raw code. However, imbalanced defect data can bias models, necessitating techniques such as data augmentation or cost-sensitive learning. Memory leaks are a common defect that can gradually deplete resources if left unrepaired. Long short-term memory (LSTM) networks are well-suited for sequential code analysis and can exploit long-term dependencies to learn complex patterns and identify leak scenarios. This research aims to advance SDP and repair through machine learning techniques.

A Tale of Two Comprehensions? Analyzing Student Programmer Attention during Code Summarization

Code summarization is the task of creating short, natural language descriptions of source code. It is an important part of code comprehension, and a powerful method of documentation. Previous work has made progress in identifying where programmers focus in code as they write their own summaries (i.e., Writing). However, there is currently a gap studying programmers’ attention as they read code with pre-written summaries (i.e., Reading). As a result, it is currently unknown how these two forms of code comprehension compare: Reading and Writing. Also, there is a limited understanding of programmer attention with respect to program semantics. We address these shortcomings with a human eye-tracking study ( n =27) comparing Reading and Writing. We examined programmers’ attention with respect to fine-grained program semantics, including their attention sequences (i.e., scan paths). We find distinctions in programmer attention across the comprehension tasks, similarities in reading patterns between them, and differences mediated by demographic factors. This can help guide code comprehension in both CS education and automated code summarization. Furthermore, we mapped programmers’ gaze data onto the Abstract Syntax Tree to explore another representation of human attention. We find that visual behavior on this structure is not always consistent with that on source code.

Big textual data research for operations management: topic modelling with grounded theory

PurposeThere is a growing need for methodological plurality in advancing operations management (OM), especially with the emergence of machine learning (ML) techniques for analysing extensive textual data. To bridge this knowledge gap, this paper introduces a new methodology that combines ML techniques with traditional qualitative approaches, aiming to reconstruct knowledge from existing publications.Design/methodology/approachIn this pragmatist-rooted abductive method where human-machine interactions analyse big data, the authors employ topic modelling (TM), an ML technique, to enable constructivist grounded theory (CGT). A four-step coding process (Raw coding, expert coding, focused coding and theory building) is deployed to strive for procedural and interpretive rigour. To demonstrate the approach, the authors collected data from an open-source professional project management (PM) website and illustrated their research design and data analysis leading to theory development.FindingsThe results show that TM significantly improves the ability of researchers to systematically investigate and interpret codes generated from large textual data, thus contributing to theory building.Originality/valueThis paper presents a novel approach that integrates an ML-based technique with human hermeneutic methods for empirical studies in OM. Using grounded theory, this method reconstructs latent knowledge from massive textual data and uncovers management phenomena hidden from published data, offering a new way for academics to develop potential theories for business and management studies.

Code observation-specific biases estimation for the modernized multi-GNSS and multi-frequency signals: An approach considering signal distortion biases

With the continuous development of global navigation satellite systems (GNSS) modernization, the code bias presented in the form of observation-specific biases (OSB) is more flexible and effective when processing modern multi-frequency and multi-GNSS signals, and therefore widely accepted by the GNSS community. However, the newly reported signal distortion biases (SDB) are ignored in the existing OSB estimation methods, which may affect the reliability of the final OSB estimation. In addition, the existing studies about SDB estimation are independent of OSB estimation, and thus a set of SDB products is generated independent of OSB products, which undoubtedly brings more complicated correction operations to GNSS users. In view of it, this study proposes an approach to code OSB estimation for multi-frequency and multi-GNSS arbitrary signals by reparameterization, which considers the SDB parameter. The aim of this study is to generate a set of code OSB products that take into account SDB parameters for the GNSS users to use. Due to considering the influence of SDB on raw code observations, the OSB estimation obtained by the proposed method is related to specific satellite-receiver pairs, which makes it different from the traditional OSB estimation methods. To achieve this, an extended model for multi-frequency and multi-GNSS arbitrary signal is first established based on the undifferenced and uncombined model, in which the SDB parameter is taken into account by re-parameterization. Based on the proposed model, three types of code bias that contain the OSB information can be obtained for the subsequent OSB estimation. Then, the three types of code bias in the first step are used as virtual observations for multi-frequency and multi-GNSS arbitrary signal OSB estimation through a two-step method. The proposed method is validated with a whole month of real measured data from 418 stations in MGEX and GA CORS to generate OSBs for GPS, GLONASS, and Galileo. The comparison between the estimated OSBs and the OSB products from CAS indicates the reliability of the proposed method. Besides, all types of OSB for four receiver groups are generated based on the proposed method to analyze the relationship between OSB and receiver groups. Results indicate that the estimated OSBs are related to receiver groups, and specific differences exist between different satellites and GNSS signals. It also proves that the OSB estimation is biased by the inconsistencies of the deviation in different receiver groups, which is more obvious for GLONASS.

A denotationally-based program logic for higher-order store

Separation logic is used to reason locally about stateful programs. State of the art program logics for higher-order store are usually built on top of untyped operational semantics, in part because traditional denotational methods have struggled to simultaneously account for general references and parametric polymorphism. The recent discovery of simple denotational semantics for general references and polymorphism in synthetic guarded domain theory has enabled us to develop TULIP, a higher-order separation logic over the typed equational theory of higher-order store for a monadic version of System F{mu,ref}. The Tulip logic differs from operationally-based program logics in two ways: predicates range over the meanings of typed terms rather than over the raw code of untyped terms, and they are automatically invariant under the equational congruence of higher-order store, which applies even underneath a binder. As a result, "pure" proof steps that conventionally require focusing the Hoare triple on an operational redex are replaced by a simple equational rewrite in Tulip. We have evaluated Tulip against standard examples involving linked lists in the heap, comparing our abstract equational reasoning with more familiar operational-style reasoning. Our main result is the soundness of Tulip, which we establish by constructing a BI-hyperdoctrine over the denotational semantics of F{mu,ref} in an impredicative version of synthetic guarded domain theory.Comment: To appear in proceedings of MFPS 2023

An Improved Doppler-Aided Smoothing Code Algorithm for BeiDou-2/BeiDou-3 un-Geostationary Earth Orbit Satellites in Consideration of Satellite Code Bias

The extensive use of carrier-aided smoothing code (CSC) filters has led to a reduction in the noise level of raw code measurements in GNSS positioning and navigation applications. However, the existing CSC technique is sensitive to the changes in the integer ambiguity, and then the smoothing procedure needs to be restarted in the presence of cycle slips. As the Doppler shift is instantaneously observed and immune to cycle slips, the Doppler-aided smoothing code (DSC) algorithm would be more promising in a challenged environment. Based on the Hatch filter, an optimal DSC approach is proposed with the principle of minimum variance. Meanwhile, to inhibit the effect of the integral cumulative error of the Doppler, a balance factor is adopted to adjust the contributions of raw code and DSC. The noise level of code observable is not only affected by thermal noise, but also limited by systematic bias. Satellite code bias (SCB) was identified in the raw code observable on each frequency for each BDS-2 satellite. By minimizing the sum of the absolute value of residuals, the polynomial segment fitting algorithm as a function of elevation angles is applied to establish the SCB correction model based on epoch-differenced multipath (MP) deviations. Finally, different types of experiments demonstrate the validity and efficiency of the refined DSC filter with SCB corrections on each available frequency for BDS un-GEO satellites.

DIRE and its Data: Neural Decompiled Variable Renamings with Respect to Software Class

The decompiler is one of the most common tools for examining executable binaries without the corresponding source code. It transforms binaries into high-level code, reversing the compilation process. Unfortunately, decompiler output is far from readable because the decompilation process is often incomplete. State-of-the-art techniques use machine learning to predict missing information like variable names. While these approaches are often able to suggest good variable names in context, no existing work examines how the selection of training data influences these machine learning models. We investigate how data provenance and the quality of training data affect performance, and how well, if at all, trained models generalize across software domains. We focus on the variable renaming problem using one such machine learning model, DIRE . We first describe DIRE in detail and the accompanying technique used to generate training data from raw code. We also evaluate DIRE ’s overall performance without respect to data quality. Next, we show how training on more popular, possibly higher quality code (measured using GitHub stars) leads to a more generalizable model because popular code tends to have more diverse variable names. Finally, we evaluate how well DIRE predicts domain-specific identifiers, propose a modification to incorporate domain information, and show that it can predict identifiers in domain-specific scenarios 23% more frequently than the original DIRE model.

IEnhancer-DCSV: Predicting enhancers and their strength based on DenseNet and improved convolutional block attention module.

Enhancers play a crucial role in controlling gene transcription and expression. Therefore, bioinformatics puts many emphases on predicting enhancers and their strength. It is vital to create quick and accurate calculating techniques because conventional biomedical tests take too long time and are too expensive. This paper proposed a new predictor called iEnhancer-DCSV built on a modified densely connected convolutional network (DenseNet) and an improved convolutional block attention module (CBAM). Coding was performed using one-hot and nucleotide chemical property (NCP). DenseNet was used to extract advanced features from raw coding. The channel attention and spatial attention modules were used to evaluate the significance of the advanced features and then input into a fully connected neural network to yield the prediction probabilities. Finally, ensemble learning was employed on the final categorization findings via voting. According to the experimental results on the test set, the first layer of enhancer recognition achieved an accuracy of 78.95%, and the Matthews correlation coefficient value was 0.5809. The second layer of enhancer strength prediction achieved an accuracy of 80.70%, and the Matthews correlation coefficient value was 0.6609. The iEnhancer-DCSV method can be found at https://github.com/leirufeng/iEnhancer-DCSV. It is easy to obtain the desired results without using the complex mathematical formulas involved.

Everyday Barriers in Communicative Participation According to People With Communication Problems.

The purpose of this study was to gain a comprehensive understanding of participation situations that are challenging for people with communication problems, to provide input for the further development of potential items for the Communicative Participation Item Bank (CPIB). A purposive sampling strategy was used to include a diverse group of people with communication problems. Diaries were used as a sensitizing exercise for inductive in-depth interviews. In these interviews, elements of communicative participation situations (concepts) were elicited that participants themselves experienced as difficult because of their communication problem. A thematic content analysis was used to identify overarching themes. In addition, new items were formulated based on the raw codes of the transcripts and linked to the International Classification of Functioning, Disability and Health Activity and Participation domains to examine the distribution of items across the breadth of the construct of communicative participation. Eighteen interviews yielded 44 different concepts. They were clustered in six themes, which capture the person, location, topic, mode, moment, and pace of communication. In total, 103 new items measuring communicative participation were formulated. Most of these items relate to International Classification of Functioning, Disability and Health Activity and Participation domains "interpersonal interactions and relationships," "major life areas," and "community, social, and civic life." This study resulted in an overview of self-reported barriers in daily communicative participation experienced by people with communication problems. These communicative participation situations can be captured within 44 concepts, which are covered by six themes. Future work should investigate if the newly written items can be added to the CPIB. The concepts and the themes can be used in designing and delivering a participation-focused intervention for this population.

Ot2Rec: A semi-automatic, extensible, multi-software tomographic reconstruction workflow.

Electron cryo-tomography (cryo-ET) is an imaging technique that produces relatively high-resolution three-dimensional images. With the advancement in microscopic and computational technologies, data collection and processing on both the hardware and software facets have become ever faster and more automated. Nowadays, tens of tilt series (up to several terabytes in total) can be collected per day, in turn shifting the experimental bottleneck to data processing. Here we present Ot2Rec, a novel workflow for performing tomographic image processing, providing access to several state-of-the-art tools within a single “wrapper” user interface. The functionality of Ot2Rec covers most tasks within a common image processing pipeline, from the motion-correction of raw multi-frame micrographs to the estimation of the true defoci from which the contrast transfer function (CTF) of images is calculated, and from the alignment of tilted image stacks to the full reconstruction of the final 3D tomogram. Moreover, Ot2Rec can carry out several novel post-processing operations such as simulation of stack point-spread functions and CTF corrections by means of deconvolution. Ot2Rec also provides automated reporting of processing parameters and quality metrics for optimisation and performance evaluation of processing workflows. Ot2Rec has been developed with the central philosophy of metadata unification. We have unified the metadata structure across the suite, so that the components within the suite can be used in a user/developer defined order (for instance, MotionCor2 -> CTFFind4 -> IMOD) with data flowing seamlessly between any pair of components and with consideration of required prerequisites. Lastly, Ot2Rec is an open-source program suite (Apache License V2). The raw code and documentation can be found at https://github.com/rosalindfranklininstitute/Ot2Rec.

Computationally efficient adaptive decompression for whole slide image processing.

Whole slide image (WSI) analysis is increasingly being adopted as an important tool in modern pathology. Recent deep learning-based methods have achieved state-of-the-art performance on WSI analysis tasks such as WSI classification, segmentation, and retrieval. However, WSI analysis requires a significant amount of computation resources and computation time due to the large dimensions of WSIs. Most of the existing analysis approaches require the complete decompression of the whole image exhaustively, which limits the practical usage of these methods, especially for deep learning-based workflows. In this paper, we present compression domain processing-based computation efficient analysis workflows for WSIs classification that can be applied to state-of-the-art WSI classification models. The approaches leverage the pyramidal magnification structure of WSI files and compression domain features that are available from the raw code stream. The methods assign different decompression depths to the patches of WSIs based on the features directly retained from compressed patches or partially decompressed patches. Patches from the low-magnification level are screened by attention-based clustering, resulting in different decompression depths assigned to the high-magnification level patches at different locations. A finer-grained selection based on compression domain features from the file code stream is applied to select further a subset of the high-magnification patches that undergo a full decompression. The resulting patches are fed to the downstream attention network for final classification. Computation efficiency is achieved by reducing unnecessary access to the high zoom level and expensive full decompression. With the number of decompressed patches reduced, the time and memory costs of downstream training and inference procedures are also significantly reduced. Our approach achieves a 7.2× overall speedup, and the memory cost is reduced by 1.1 orders of magnitudes, while the resulting model accuracy is comparable to the original workflow.

Coupling a weather model directly to GNSS orbit determination – case studies with OpenIFS

Abstract. Neutral gas atmosphere bends and delays propagation of microwave signals in satellite-based navigation. Weather prediction models can be used to estimate these effects by providing three-dimensional refraction fields to ray-trace the signal delays. In this study, a global numerical weather prediction model (Open Integrated Forecasting System (OpenIFS) licensed for Academic use by the European Centre for Medium-Range Weather Forecast) is used to generate the refraction fields. The ray-traced slant delays are supplied as such – in contrast to mapping – for an orbit solver (GROOPS (Gravity Recovery Object Oriented Programming System) software toolkit of Graz University of Technology) which applies the raw observation method. Here we show that such a close coupling is possible without need for major additional modifications in the solver codes. The main finding here is that the adopted approach provides a very good a priori model for the atmospheric effects on navigation signals. We suspect that removal of the intermediate mapping step allows us to take advantage of the local refraction field asymmetries in the GNSS signal processing. Moreover, the direct coupling helps in identifying deficiencies in the slant delay computation because the modeling errors are not convoluted in the mapping procedures. These conclusions appear robust, despite the relatively small data set of raw code and phase observations covering the core network of 66 ground-based stations of the International GNSS Service over 1-month periods in December 2016 and June 2017. More generally, the new configuration enhances our control of geodetic and meteorological aspects of the orbit problem. This is pleasant because we can, for instance, regulate at will the weather model output frequency and increase coverage of spatiotemporal aspects of weather variations. The direct coupling of a weather model in precise GNSS orbit determination presented in this paper provides a unique framework for benefiting even more widely than previously the apparent synergies in space geodesy and meteorology.

HELP-DKT: an interpretable cognitive model of how students learn programming based on deep knowledge tracing

Student cognitive models are playing an essential role in intelligent online tutoring for programming courses. These models capture students’ learning interactions and store them in the form of a set of binary responses, thereby failing to utilize rich educational information in the learning process. Moreover, the recent development of these models has been focused on improving the prediction performance and tended to adopt deep neural networks in building the end-to-end prediction frameworks. Although this approach can provide an improved prediction performance, it may also cause difficulties in interpreting the student’s learning status, which is crucial for providing personalized educational feedback. To address this problem, this paper provides an interpretable cognitive model named HELP-DKT, which can infer how students learn programming based on deep knowledge tracing. HELP-DKT has two major advantages. First, it implements a feature-rich input layer, where the raw codes of students are encoded to vector representations, and the error classifications as concept indicators are incorporated. Second, it can infer meaningful estimation of student abilities while reliably predicting future performance. The experiments confirm that HELP-DKT can achieve good prediction performance and present reasonable interpretability of student skills improvement. In practice, HELP-DKT can personalize the learning experience of novice learners.

Relative Positioning in Remote Areas Using a GNSS Dual Frequency Smartphone.

The use of GPS positioning and navigation capabilities in mobile phones is present in our daily lives for more than a decade, but never with the centimeter level of precision that can actually be reached with several of the most recent smartphones. The introduction of the new GNSS systems (Global Navigation Satellite Systems), the European system Galileo, is opening new horizons in a wide range of areas that rely on precise georeferencing, namely the mass market smartphones apps. The constant growth of this market has brought new devices with innovative capabilities in hardware and software. The introduction of the Android 7 by Google, allowing access to the GNSS raw code and phase measurements, and the arrival of the new chip from Broadcom BCM47755 providing dual frequency in some smartphones came to revolutionize the positioning performance of these devices as never seen before. The Xiaomi Mi8 was the first smartphone to combine those features, and it is the device used in this work. It is well known that it is possible to obtain centimeter accuracy with this kind of device in relative static positioning mode with distances to a reference station up to a few tens of kilometers, which we also confirm in this paper. However, the main purpose of this work is to show that we can also get good positioning accuracy using long baselines. We used the ability of the Xiaomi Mi8 to get dual frequency code and phase raw measurements from the Galileo and GPS systems, to do relative static positioning in post-processing mode using wide baselines, of more than 100 km, to perform precise surveys. The results obtained were quite interesting with RMSE below 30 cm, showing that this type of smartphone can be easily used as a low-cost device, for georeferencing and mapping applications. This can be quite useful in remote areas where the CORS networks are not dense or even not available.

Automated characterisation of glaciomarine sediments using X-ray computed laminography

This study investigates the potential of a new high-resolution, non-destructive, X-ray imaging technique for the Quaternary Sciences – computed laminography (CL). Greyscale properties are systematically extracted from digital X-radiographic CL images of cored glaciomarine sediments to analyse and characterise sediments at pixel-scale resolution (<0.1 mm). We show how this can be achieved manually, and also with an easy-to-use, automated statistical tool which we have devised specifically for use in glaciomarine sediments. This Sediment Characteristics tool, in the form of a plugin for the freely available FIJI/ImageJ programme, extracts mean or median X-ray grey values (GV) – a proxy for sediment density; and associated standard deviation (SD) – a proxy for sediment heterogeneity – at sub-mm resolution, across the width of sediment core CL images. We demonstrate how these properties (GV and SD) can be directly used to characterise sediment properties and in particular to quantify the abundance of gravel clasts, or ice-rafted debris, in cored glaciomarine sediments. The tool’s effectiveness is compared with other, more traditional, X-radiographic methods for counting ice-rafted gravel clasts in glaciomarine sediment. We propose that the CL output and Sediment Characteristics tool also have the potential to quantitatively analyse other 3-dimensional structures, such as cyclic lamination (varve) geometry; deformation structures; bioturbation and void space (porosity). Finally, we present the raw code, allowing open-access, transparency and reproducibility in other formats.

Analysis of Finger Print Bank Algorithm based Fingerprint Matching Scheme

Now-a-days it is an important need to preserve one’s data/information securely without giving an option to attackers or intruders to steal it. For this case authentication are considered to be the most important feature, which allows everyone to individually prove their authentication to the system and once it is correct allow them to access the features of the system otherwise neglect or block them to proceed further. There are many authentication schemes available in the information technology and security industries such as Biometric Scheme, IRIS Matching Scheme, Face Recognition, Password Management Scheme and so on. The most popular, well-known and one of the best classical scheme is called Biometric Finger Print Matching Scheme, which allows the user to register the finger print into the system for training purpose and further at every time of accessing the features into the system it gathers the present (testing) finger print from the user and match it with the already registered finger print, once it matches with the trained sample, it allows the user to login into the system or else block the user to proceed further. The proposed Finger Print Bank Algorithm uses an efficient finger print matching principles to accurately match the correct finger print and gives a Boolean result to user to inform that to proceed further or not. The Finger Print Bank Algorithm uses efficient filtering schemes to filter the finger print to extract the internal and global core details of it as well as extracts the raw code of it and compares that with the already registered finger print. For all the entire proposed approach guarantees for the best result and accuracy in results.

Conditioning and PPP processing of smartphone GNSS measurements in realistic environments

Smartphones typically compute position using duty-cycled Global Navigation Satellite System (GNSS) L1 code measurements and Single Point Positioning (SPP) processing with the aid of cellular and other measurements. This internal positioning solution has an accuracy of several tens to hundreds of meters in realistic environments (handheld, vehicle dashboard, suburban, urban forested, etc.). With the advent of multi-constellation, dual-frequency GNSS chips in smartphones, along with the ability to extract raw code and carrier-phase measurements, it is possible to use Precise Point Positioning (PPP) to improve positioning without any additional equipment. This research analyses GNSS measurement quality parameters from a Xiaomi MI 8 dual-frequency smartphone in varied, realistic environments. In such environments, the system suffers from frequent phase loss-of-lock leading to data gaps. The smartphone measurements have low and irregular carrier-to-noise (C/N0) density ratio and high multipath, which leads to poor or no positioning solution. These problems are addressed by implementing a prediction technique for data gaps and a C/N0-based stochastic model for assigning realistic a priori weights to the observables in the PPP processing engine. Using these conditioning techniques, there is a 64% decrease in the horizontal positioning Root Mean Square (RMS) error and 100% positioning solution availability in sub-urban environments tested. The horizontal and 3D RMS were 20 cm and 30 cm respectively in a static open-sky environment and the horizontal RMS for the realistic kinematic scenario was 7 m with the phone on the dashboard of the car, using the SwiftNav Piksi Real-Time Kinematic (RTK) solution as reference. The PPP solution, computed using the YorkU PPP engine, also had a 5–10% percentage point more availability than the RTK solution, computed using RTKLIB software, since missing measurements in the logged file cause epoch rejection and a non-continuous solution, a problem which is solved by prediction for the PPP solution. The internal unaided positioning solution of the phone obtained from the logged NMEA (The National Marine Electronics Association) file was computed using point positioning with the aid of measurements from internal sensors. The PPP solution was 80% more accurate than the internal solution which had periodic drifts due to non-continuous computation of solution.

Impact of diagnosis code grouping method on clinical prediction model performance: A multi-site retrospective observational study

ObjectiveThe primary purpose of this work is to systematically assess the performance trade-offs on clinical prediction tasks of four diagnosis code groupings: AHRQ-Elixhauser, Single-level CCS, truncated ICD-9-CM codes, and raw ICD-9-CM codes. Materials and methodsWe used two distinct datasets from different geographic regions and patient populations and train models for three prediction tasks: 1-year mortality following an ICU stay, 30-day mortality following surgery, and 30-day complication following surgery. We run multiple commonly-used binary classification models including penalized logistic regression, random forest, and gradient boosted trees. Model performance is evaluated using the Area Under the Receiver Operating Characteristic (AUROC) and the Area Under the Precision-Recall Curve (AUCPR). ResultsSingle-level CCS, truncated codes, and raw codes significantly outperformed AHRQ-Elixhauser ICD grouping when predicting 30-day postoperative complication and one-year mortality after ICU admission. The performance across groupings was more similar in the 30-day postoperative mortality prediction task. DiscussionSingle-level CCS groupings represent aggregations of raw codes into meaningful clinical concepts and consistently balance interoperability between ICD-9-CM and ICD-10-CM while maintaining strong model performance as measured by AUROC and AUCPR. Key limitations include experimentation across two datasets and three prediction tasks, which although were well labeled and sufficiently prevalent, do not encompass all modeling tasks and outcomes. ConclusionSingle-level CCS groupings may serve as a good baseline for future models that incorporate diagnosis codes as features in clinical prediction tasks. Code and a compute environment summary are provided along with the analyses to enable reproducibility and to support future research.