Essays about: "Outlier removal"
Showing result 1 - 5 of 7 essays containing the words Outlier removal.
-
1. Comparing the effect of random and contextual removal of images on object detection performance
University essay from Högskolan i Skövde/Institutionen för informationsteknologiAbstract : As datasets grow, the need for automated methods to ensure dataset quality arises. This report presents an experiment conducted on the MSCOCO train2017 dataset to identify image outliers using a force-directed graph built from a co-occurrence context, focusing on the mean average precision and average precision. READ MORE
-
2. Modelling the full-scale N2O emissions from wastewater treatment plants for identifying mitigation strategies
University essay from Lunds universitet/Kemiteknik (CI)Abstract : Nitrous oxide (N2O) emissions, which primarily originate from the biological nitrogen removal process, dominate the release of greenhouse gases (GHGs) in wastewater treatment plants (WWTPs). N2O production occurs through dynamic microbial pathways that have a significant impact on the environment compared with other GHG emissions. READ MORE
-
3. LiDAR-Equipped Wireless Sensor Network for Speed Detection on Classification Yards
University essay from KTH/Skolan för elektroteknik och datavetenskap (EECS)Abstract : Varje dag kopplas tusentals godsvagnar om på de olika rangerbangårdarna i Sverige. För att kunna automatiskt bromsa vagnarna tillräckligt mycket är det nödvändigt att veta deras hastigheter. En teknik som har blivit populär på sistone är Light Detection and Ranging (LiDAR) som använder ljus för att mäta avstånd till objekt. READ MORE
-
4. Maps element prediction from raw time-series position data using machine learning techniques
University essay from Uppsala universitet/Institutionen för informationsteknologiAbstract : Having geopositional tracking data of users moving in the work space, we would like to detect map elements such as paths, zones of interest and borders and to create abroad map of the working area from this time series raw data. In order to do so, an experiment to generate data was conducted to emulate workers moving around thework-space. READ MORE
-
5. Automated error matching system using machine learning and data clustering : Evaluating unsupervised learning methods for categorizing error types, capturing bugs, and detecting outliers.
University essay from Linköpings universitet/Programvara och systemAbstract : For large and complex software systems, it is a time-consuming process to manually inspect error logs produced from the test suites of such systems. Whether it is for identifyingabnormal faults, or finding bugs; it is a process that limits development progress, and requires experience. READ MORE