Machine training

AI & Machine Learning, Python — incorrectly specified categories?

45 USD

Project translated automatically. Log in or register, to view the original The work must be done in the jupiter notebook + explanatory note(can be in the form of comments and explanations in jupiter ) Subject - "Application of machinery for the promotion of tourist yslygi" The specific area in the tourist services to which the MO will be applied, is selected by the executive independently.Based on this topic, the following points should be fulfilled: The !Points 1,3,11,9 are mandatory for all the works!1 .Select (collect) data set for analysis according to the chosen course theme The work.Describe this set and the task to be solved.Determine the quality metrics for the model, to determine the typical level of metric to solve similar tasks, to establish Guidelines for Metrics.2ndPre-analysis and data cleaning.This phase includes the conclusion. information about quantitative characteristics of the data network, information about the missing the meanings, characteristics and physical meaning of each attribute of data, its significance for the prediction of the target variable, the output of several data elements for Illustration of the data structure.ThreeIf necessary, perform the data conversion.This stage depends greatly on the type. the data being investigated and may include tokinization and vectorization of the text, extracting signs from data, converting images into a flat number mass and other transformations.4 .Divide the data set into training, test and validation samples.to justify Quantitative characteristics and method of separation (temporary, random, The consistent samples.and 5.Defining the principle architecture of the neuron network model of deep learning, Includes the error function and regulation methods, the general logic of model learning.Predicts possible variations of model modification.6 .Implement the model using the PyTorch framework.Implementation must include the functionality of data loading, the removal of intermediate and final quality meters and Diagnostic parameters of the model learning process.Implementation must provide flexibility to enable limited modification of the model Choosing the best of alternative architectures.7 .Set the input data parameters and the learning parameters.to perform training Models on prepared input data using selected settings and algorithms and training.8 .To evaluate the quality of the model using the quality metrics provided, Check the model for re-learning.The 9.Provides training and quality assessment for several alternative architectures (variations) architecture) models (at least 3) to solve the chosen task.Analysis Results, making conclusions, including determining the best architecture.The 10.Implement a thin setting of the model by selecting the values of hyperparameters.Select at least three hyperparameters as a minimum for one selection. The parameters are performed in automatic mode (with the Grid Search or the similar methods).11 of 11.Present the learning results of the model in a clear form (graphics, learning lines, comparison tables of models, classification tables, and others).To make conclusions, to compare with the existing similar solutions, to raise the prospects of improvement Methods of solving the problem.Depending on the formulation of the chosen topic, the volume and the presence of points in this list can vary.Points 1,3,11,9 are mandatory for all works.

Proposals 3