Sökning: "time series classification"
Visar resultat 21 - 25 av
136 uppsatser innehållade orden time series classification.
-
Master-uppsats, Stockholms universitet/Institutionen för data- och systemvetenskap
Författare :Markus Schultz; [2022]
Nyckelord :Time Series; Counterfactual; Model Agnostic; Machine Learning; Classification;
Sammanfattning :
Time Series are used in healthcare, meteorology, and many other fields. Rigorous research has been done to develop distance measures and classifying algorithms for time series. When a time series is classified, one can ask what changes should be made to the time series to classify it differently. LÄS MER
-
-
Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)
Författare :Farid Hedayati; [2022]
Nyckelord :Magnetospheric Multiscale Mission MMS ; Magnetosheath classification; Particles; Time series analysis; Space plasma physics;
Sammanfattning :
The emphasis of the thesis is on understanding how the high-energy particles of the foreshock are transmitted into the magnetosheath and how this process can be used to specify the conditions upstream of the bow shock with the help of local magnetosheath measurements. In this study, through analyzing the data from the MMS satellites, I try to draw a connection between the flux and wind direction related to Earth’s bow shock. LÄS MER
-
Master-uppsats, Uppsala universitet/Luft-, vatten- och landskapslära
Författare :Arnór Tumi Jóhannsson; [2022]
Nyckelord :Wind power; Iceland; Atmospheric Boundary Layer; atmospheric stability; probabilistic forecasting; Machine Learning; Quantile Regression Forest; Vindkraft; Island; gränsskikt; atmosfärisk stabilitet; probabilistisk prognos; Machine Learning; Quantile Regression Forest;
Sammanfattning :
Accurate forecasts of wind speed and power production are of great value for wind power producers. In Southwest Iceland, wind power installations are being planned by various entities. LÄS MER
-
Master-uppsats, Linköpings universitet/Statistik och maskininlärning
Författare :priya kurian pullolickal; [2022]
Nyckelord :Electrocardiogram ECG measures the electrical impulses of the heart. The time inter- val between two successive R peaks measured in millisecond using an ECG is called as an RR-interval. The distribution of the RR-intervals as well as classification of cardiovascu- lar age of healthy people from RR-interval was done in this thesis. For that; the data was preprocessed and time series plots were analyzed from sample dataset. The RR-intervals were then aligned to have the same start time using functions written and then an aver- age RR-interval series for each decade was created. The coefficient of variation was very less for this averaged dataset which concluded that averaging the RR-interval was a good approach. The averaged dataset per age decade as well agreed to the conclusion of the sample data set that the heart rate variability decreases with increasing age. Three clusters of age decade were also visible in the averaged dataset. The kurtosis; skew; mean; me- dian; histograms and Q-Q Plot were calculated for the sample as well as averaged dataset to find the distribution. The values all concluded that the RR-intervals follow Gaussian distribution or mixture of Gaussian distribution. The Poincaré plots showed that the dis- tribution of RR-interval is comet shaped for healthy individuals. The features were ex- tracted from the distribution as well as from the distribution of Discrete Fourier Transform DFT for classifying the age group from RR-intervals. Svitzky-Golay filtering was done to smooth the signal before taking the features from DFT. Random Forest and Support Vector Machine was the machine learning algorithms used to classify the age decade. Later the results were compared using a dataset from physionet that had RR-intervals of individuals suffering from myocardial infraction. The age classification using Random Forest and Sup- port Vector Machine concluded that the Gdańsk dataset using Random Forest Algorithm and three classes gave the highest test accuracy of 59% for the dataset.;
Sammanfattning :
.... LÄS MER
-
-
Master-uppsats, Lunds universitet/Produktionsekonomi
Författare :Artur Jusopov; Arian Marofkhani; [2022]
Nyckelord :Forecasting; Intermittent Demand; Spare Parts; After-Market; Crostons; Exponential Smoothing; Hierarchical Forecasting; Technology and Engineering;
Sammanfattning :
Title: After-Market Spare Parts Forecasting at Sandvik Stationary Crushing & Screening. Authors: Arian Marofkhani and Artur Jusopov Supervisors: Professor Gudrun Kiesmüller, Lund University, Faculty of Engineering, Division of Production Management. Macarena Ribalta, Planning & Logistics Manager, Sandvik Stationary Crushing & Screening. LÄS MER
-
