from ModelMain import s_squared, improved_gradient_descent_step from QuickDataVisualizer import write_comparison_csv from Utils import get_data_from_csv, find_t0_data_point from DifferentialModels import Harissons from AnalyisisMain import fetch_parameters # create an instance of the EP model dt = 0.005 # parameters for the Euler's method algorithm max_t = 6.6 model = Harissons(dt, max_t) parameters = [fetch_parameters("../HarissonsAnalysisResults", False, data_set_num=2, step_num=1853, method="AGD")] # some parameters e1 = 0.01 # value of e1 used in the calculating S^2 e2 = 0.025 # value of e2 used in the calculating S^2 datas = [get_data_from_csv("../Data/exp4_beker{}_data.csv".format(i), t_scale=1) for i in range(1, 5)] step = improved_gradient_descent_step([datas[1]], model, parameters[0], 1e-5, 1e-1, minimal_second_gradient=1e-7, e1=0.01, e2=0.025) parameters.append(parameters[0].copy()) for key in parameters[0].keys(): parameters[-1][key] = parameters[0][key] + step[key] for p in parameters: print(p) write_comparison_csv("test_result7.csv", datas[1], model, parameters, 0.01)
from ModelMain import s_squared, improved_gradient_descent_step from QuickDataVisualizer import write_comparison_csv from Utils import get_data_from_csv from DifferentialModels import CLV from AnalyisisMain import fetch_parameters # create an instance of the EP model dt = 0.005 # parameters for the Euler's method algorithm max_t = 6.6 model = CLV(dt, max_t) parameters = [fetch_parameters("CLVAnalysisResults", False, data_set_num=3, step_num=10874+i, method="AGD") for i in range(-4, 3)] # some parameters e1 = 0.01 # value of e1 used in the calculating S^2 e2 = 0.025 # value of e2 used in the calculating S^2 datas = [get_data_from_csv("Data/exp4_beker{}_data.csv".format(i), t_scale=1) for i in range(1, 5)] for p in parameters: print(s_squared(datas[2], model, p, e1, e2)) for p in parameters: print(improved_gradient_descent_step([datas[2]], model, p, 1e-5, 2e-1, minimal_second_gradient=1e-7, e1=0.01, e2=0.025)) write_comparison_csv("../test_result.csv", datas[2], model, parameters, 0.01)
from ModelMain import s_squared, improved_gradient_descent_step from QuickDataVisualizer import write_comparison_csv from Utils import get_data_from_csv, find_t0_data_point from DifferentialModels import Harissons from AnalyisisMain import fetch_parameters # create an instance of the EP model dt = 0.005 # parameters for the Euler's method algorithm max_t = 6.6 model = Harissons(dt, max_t) parameters = [fetch_parameters("../HarissonsAnalysisResults", True, step_num=192, method="AGD")] # some parameters e1 = 0.01 # value of e1 used in the calculating S^2 e2 = 0.025 # value of e2 used in the calculating S^2 datas = [get_data_from_csv("../Data/exp4_beker{}_data.csv".format(i), t_scale=1) for i in range(1, 5)] step = improved_gradient_descent_step(datas, model, parameters[0], 1e-5, 1e-1, minimal_second_gradient=1e-7, e1=0.01, e2=0.025) parameters.append(parameters[0].copy()) for key in parameters[0].keys(): parameters[-1][key] = parameters[0][key] + step[key] for p in parameters: print(p) write_comparison_csv("test_result8.csv", datas[2], model, parameters, 0.01)
] # create an instance of the CLV model dt = 0.005 # parameters for the Euler's method algorithm max_t = 6.6 model = CLV(dt, max_t) # stuff min_step = 10874 m0 = 1e-7 m_factor = 10**(1 / 10) f_min = -30 f_max = 50 data_set_num = 3 parameters = fetch_parameters("../CLVAnalysisResults", False, min_step, "AGD", data_set_num) errors = [] for f in range(f_min, f_max + 1): m = m0 * (m_factor**f) step = improved_gradient_descent_step([datas[data_set_num - 1]], model, parameters, epsilon, step_size, minimal_second_gradient=m, e1=e1, e2=e2) p = parameters.copy() for key in p.keys():