32584 research outputs found
Sort by
Assessing GTFS Accuracy
Get PDFThe promised benefits of the General Transit Feed Specification (GTFS) Schedule and Realtime standards are dependent on the underlying quality of the data. Despite this fundamental reliance, there has been relatively little research on techniques and strategies to assess GTFS accuracy. The need for such assessment is growing as federal and state governments increasingly require transit agencies to make these data available to the public. This research fills this gap by presenting a suite of methods and metrics to assess the temporal accuracy of GTFS Realtime and the spatial accuracy of GTFS Schedule feeds. The temporal assessment demonstrates an approach to collect and clean TripUpdate messages to identify (and derive) a set of values for measuring the accuracy of the vehicle arrival predictions. These metrics are carefully designed to provide transit agencies insight into the quality of the data they provide to customers in terms of the impact of those inaccuracies on the customer experience. The spatial assessment demonstrates an approach to match scheduled information on the location of transit routes and stops with the actual travel patterns demonstrated in the realtime VehiclePosition messages. The measured divergence between the planned and provided transit service yields a series of location accuracy metrics. All of the proposed metrics can be scaled to examine GTFS accuracy from the stop to the systemwide level. All of the proposed metrics can be easily generated from publicly available GTFS feeds without any additional data sources. Finally, all of the proposed metrics can help transit agencies continuously assess and therefore improve the quality of GTFS data they share with the public
Forecasting Commercial Vehicle Miles Traveled (VMT) in Urban California Areas
Get PDFThis study investigates commercial truck vehicle miles traveled (VMT) across six diverse California counties from 2000 to 2020. The counties—Imperial, Los Angeles, Riverside, San Bernardino, San Diego, and San Francisco—represent a broad spectrum of California’s demographics, economies, and landscapes. Using a rich dataset spanning demographics, economics, and pollution variables, we aim to understand the factors influencing commercial VMT. We first visually represent the geographic distribution of the counties, highlighting their unique characteristics. Linear regression models, particularly the least absolute shrinkage and selection operator (LASSO) and elastic net regressions are employed to identify key predictors of total commercial VMT. LASSO regression emphasizes feature selection, revealing vehicle population and fuel consumption as significant predictors in most counties. Elastic net regression, which balances feature selection and multicollinearity, expands the list of predictors to include variables like the number of trips, CO2 emissions, and PM2.5 pollution. Overall, the findings suggest that economic factors, such as fuel consumption and vehicle population, significantly impact the total commercial VMT across the counties. Pollution variables, specifically CO2 and PM2.5, also play a role. These insights underscore the need for nuanced transportation and environmental policies, especially in the face of economic fluctuations, to manage commercial truck VMT effectively and sustainably. Methodology using both LASSO and elastic net regression provides a robust framework for understanding these complex relationships in commercial transportation behavior
Spartan Daily, September 12, 2024
Get PDFVolume 163, Issue 9https://scholarworks.sjsu.edu/spartan_daily_2024/1049/thumbnail.jp
Governing Structures for Successful Regional Transit Coordination and their Formation
Get PDFThe expansion of metropolitan areas in California has further prompted the need to improve transportation options for all, more effectively linking origins and destinations through key enhancements to the existing network of transit services. This study provides planners and policymakers examples of effective regional transit coordination agencies. To improve multimodal connections and enhance transit services at the local and regional levels, this study explores regional coordination, focusing on entities charged with coordinating multiple transit agencies in a single metropolitan area. The study identified 16 regional transit coordinators (RTCs), identifying the structure, scope, and management of each. The results revealed that there are many factors involved in creating an organization with the authority to coordinate regional transit, including how the organization was established and elements related to the formation of an RTC (e.g., regional dynamics and board composition). It is also important to study the powers vested in different types of boards and members, as well as the executive director. There is a wide range of state legislation for the legal establishment of RTCs. Voluntary transit federations, another option, fall into two categories: loose federations, based on consensus and strong federations, based on binding arrangements to coordinate fares, services, and information. The authors offer options for establishing an RTC in California, including a multi-county owned corporation with an ex officio board; an agency of the state or regional government; and a special district with an ex officio board representing counties, cities, planning staff, operator staff, and state transport agencies. Finally, suggestions for creating the board are provided along with a recommendation that the state develop a legislative framework to establish RTCs in its metropolitan areas
A Two-stage Machine Learning Approach for Fake News Detection and News Article Categorization
Get PDFThis project employs machine learning techniques to develop a sequential model for detecting and categorizing fake news, aiming to mitigate its proliferation in today\u27s digital landscape. The model operates in two phases: in the first phase, the classification algorithms like Naïve Bayes, XGBoost and Random Forest are used to distinguish between true and false news stories and in the second phase the capabilities of Naïve Bayes, XGBoost, Random Forest, and the Transformer-based BERT (Bidirectional Encoder Representations from Transformers) model are leveraged to further categorize the news into specific topics.
The methodology encompasses several key steps: data acquisition, preprocessing, feature extraction, and model training, followed by evaluation using metrics such as accuracy, F1 score and a detailed classification report. These processes ensure that the model not only identifies fake news but also classifies it effectively alongside legitimate articles
Characterizing Nanopore Sequencing Artifacts with Deep Learning
Get PDFOxford Nanopore sequencing is a revolutionary new technology for sequencing DNA molecules in long stretches. However, it has a significantly higher error rate than conventional short-read sequencing, resulting in numerous sequencing artifacts. These artifacts can be indistinguishable from low frequency somatic variants, which is a roadblock for cancer diagnosis using liquid biopsies. In this study, benchmarked human genome samples from Genome in a Bottle were used to create a dataset of labeled variants, including artifacts and true variants. Variant features, including sequence context, were used to train various deep learning models. The multi-input neural network combining sequence context features and other variant features resulted in higher validation accuracy (0.871) than the other non-sequence context features alone (0.853), demonstrating that the sequence context surrounding a variant has some predictive power regarding whether a called variant is a sequencing artifact or a true variant
An at-home evaluation of a light intervention to mitigate sleep inertia symptoms
Get PDFObjectives: Under laboratory settings, light exposure upon waking at night improves sleep inertia symptoms. We investigated whether a field-deployable light source would mitigate sleep inertia in a real-world setting. Methods: Thirty-six participants (18 female; 26.6 years ± 6.1) completed an at-home, within-subject, randomized crossover study. Participants were awoken 45 minutes after bedtime and wore light-emitting glasses with the light either on (light condition) or off (control). A visual 5-minute psychomotor vigilance task, Karolinska sleepiness scale, alertness and mood scales, and a 3-minute auditory/verbal descending subtraction task were performed at 2, 12, 22, and 32 minutes after awakening. Participants then went back to sleep and were awoken after 45 minutes for the opposite condition. A series of mixed-effect models were performed with fixed effects of test bout, condition, test bout × condition, a random effect of the participant, and relevant covariates. Results: Participants rated themselves as more alert (p = .01) and energetic (p = .001) in the light condition compared to the control condition. There was no effect of condition for descending subtraction task outcomes when including all participants, but there was a significant improvement in descending subtraction task total responses in the light condition in the subset of participants waking from N3 (p = .03). There was a significant effect of condition for psychomotor vigilance task outcomes, with faster responses (p \u3c .001) and fewer lapses (p \u3c .001) in the control condition. Conclusions: Our findings suggest that light modestly improves self-rated alertness and energy after waking at home regardless of sleep stage, with lower aggression and improvements to working memory only after waking from N3. Contrary to laboratory studies, we did not observe improved performance on the psychomotor vigilance task. Future studies should include measures of visual acuity and comfort to assess the feasibility of interventions in real-world settings
Optimizing Field-Linked Simulations of Dry Season Uptake and Monsoon Infiltration within an Aspen-Mixed Conifer Forest
Get PDFClimate models forecast that headwater catchments in the western U.S. will undergo a reduction in snowpack, early season snowmelt, and increases in evapotranspiration. The resulting extended dry season will stress vegetation in mountainous watersheds throughout the Upper Colorado River Basin. We investigate infiltration patterns and root water uptake in response to dry season disturbances within the East River watershed in Colorado. To do this, we collected soil cores, measured matric potential and sap flow, and monitored tree xylem and soil for stable isotopes of water (2H, 18O) in two soil profiles to 90 cm depth (with three Engelmann spruce and three aspen trees instrumented, respectively). Sub-daily stable water isotope dynamics were analyzed between mid-June and late October of WY-22 using a cavity ring-down spectrometer. HYDRUS-1D was calibrated with both field measurements of matric potential and δ2H to simulate the ecosystem response to late summer dry spells and monsoonal rainfall for the 128-day time domain. Simulations show both aspen and spruce trees relying heavily on headwater snowmelt. Water-use analysis reveals a reduction in transpiration during the late summer and early autumn months, particularly in September, the driest month of WY-22. Spruce trees appear more tolerant of dry soil conditions than aspen trees. As a result, it is likely that spruce will outcompete aspen in the future, given forecasted shifts in climate and changes in the timing of snowmelt. This study highlights the benefits of coupling high frequency stable water isotope data and more commonly used matric potential field measurements for parameter optimization in numerical modeling
Moore, Robert J. (1923-2024)
No full textUniversity of Buffalo (NY), 1956 Ed.D.
University of Minnesota, 1948 MA
University of Minnesota, 1945 BShttps://scholarworks.sjsu.edu/erfa_bios/1192/thumbnail.jp