|
|
Machine Learning Optimization of KPI Prediction
Haris, Daniel ; Burget, Radek (referee) ; Bartík, Vladimír (advisor)
This thesis aims to optimize the machine learning algorithms for predicting KPI metrics for an organization. The organization is predicting whether projects meet planned deadlines of the last phase of development process using machine learning. The work focuses on the analysis of prediction models and sets the goal of selecting new candidate models for the prediction system. We have implemented a system that automatically selects the best feature variables for learning. Trained models were evaluated by several performance metrics and the best candidates were chosen for the prediction. Candidate models achieved higher accuracy, which means, that the prediction system provides more reliable responses. We suggested other improvements that could increase the accuracy of the forecast.
|
|
|
The characterization of the collimated beams of fast neutrons with the CLID detecion system
Ansorge, Martin ; Novák, Jan ; Majerle, Mitja ; Kozic, Ján
A new detection device for the measurements of light ions (p, d, t, α) emitted as the products of the nuclear reactions induced by fast neutrons (5-33 MeV) was recently developed at the Nuclear Physics Institute of the Czech Academy of Sciences. The main objective of the Chamber-for-Light-Ion-Detection (CLID) is to produce new differential nuclear data of high interest for the material applications related to fusion and aerospace technologies and to potentially test and validate models of nuclear reactions. Hereby the experimental set-up for the measurements with the CLID is described in detail. The experimental characterization of the collimated fast neutron beams produced by the cyclotron-driven converter (p(35 MeV)+Be(2.5 mm)) is presented. In particular, the implementation of the Proton-Recoil-Telescope technique used for neutron energy spectra determination with the CLID is described.
|
|
|
Machine Learning Optimization of KPI Prediction
Haris, Daniel ; Burget, Radek (referee) ; Bartík, Vladimír (advisor)
This thesis aims to optimize the machine learning algorithms for predicting KPI metrics for an organization. The organization is predicting whether projects meet planned deadlines of the last phase of development process using machine learning. The work focuses on the analysis of prediction models and sets the goal of selecting new candidate models for the prediction system. We have implemented a system that automatically selects the best feature variables for learning. Trained models were evaluated by several performance metrics and the best candidates were chosen for the prediction. Candidate models achieved higher accuracy, which means, that the prediction system provides more reliable responses. We suggested other improvements that could increase the accuracy of the forecast.
|
|
|
Measurements and usage of cross sections of various (n,chi n) threshold reactions
Chudoba, Petr ; Vrzalová, Jitka ; Svoboda, Ondřej ; Krása, Antonín ; Kugler, Andrej ; Majerle, Mitja ; Suchopár, Martin ; Wagner, Vladimír
Current trend in nuclear reactor physics is a transition from technologies using thermal neutrons to technologies utilizing fast neutrons. Unfortunately focus was put mainly on the thermal neutrons for a long time and lead to very good knowledge about this low energy region, but very scarce coverage of the high energy region. This means that there is a gap in the knowledge of excitation functions for higher energies. This gap spreads from 20 MeV up to 1 GeV and higher. This is exactly the energy region needed for description of advanced nuclear systems such as accelerator driven systems (ADS). Our group from Nuclear Physics Institute (NPI) of the CAS is a member of an international collaboration Energy & Transmutation of Radioactive Waste (E&T RAW). This collaboration focuses on ADS for many years. In order to measure neutron field within ADS models it is necessary to know excitation functions of reactions used to monitor the neutron field. In many cases there are almost no experimental data for suitable reactions. Worse and quite common case is that there are no data at all. Therefore we are also focusing on measurements of these data in order to fill the databases as well as to allow further improvements of codes for nuclear data calculations.
|
guest ::
