CVSSForecaster: Class CVSSForecaster

In realerikrani/nvdr: NVD Vulnerability Data Analyser

Description

Forecasts monthly mean CVSS scores of subsets of CWE vulnerabilities.

Private attributes

All private attribute's have NA as default value.

• cwes the character vector of available CWEs

• cwes_ts monthly time series of mean monthly CVSS scores of the CWEs in private attribute cwes

• start_year the first year of a time period

• end_year the last year of a time period

• end_month the last month of a time period

• benchmark the attribute for holding a list of BenchmarkModel objects

• arima the attribute for holding a list of ARIMAModel objects

• ets the attribute for holding a list of ETSModel objects

• tslinear the attribute for holding a list of TSLinearModel objects

• nnar the attribute for holding a list of NNARModel objects

• arfima the attribute for holding a list of ARFIMAModel objects

• bagged_ets the attribute for holding a list of BaggedETSModel objects

• tbats the attribute for holding a list of TBATSModel objects

• struct_ts the attribute for holding a list of StructTSModel objects

• best_model_list the attribute for holding a list of forecasting model objects which gave the best overall forecast accuracy results for specific CWEs

Public methods

• new() creates new CVSSForecaster object

• getCWENames() returns private attribute cwes

• getSeries() returns private attribute cwes_ts

• getBenchmark(cwe_name) returns private attribute benchmark when the argument cwe_name is missing; otherwise returns a BenchmarkModel object (from benchmark) corresponding to cwe_name value (CWE ID)

• getARIMA(cwe_name) returns private attribute arima when the argument cwe_name is missing; otherwise returns a ARIMAModel object (from arima) corresponding to cwe_name value (CWE ID)

• getETS(cwe_name) returns private attribute ets when the argument cwe_name is missing; otherwise returns a ETSModel object (from ets) corresponding to cwe_name value (CWE ID)

• getTSLinear(cwe_name) returns private attribute tslinear when the argument cwe_name is missing; otherwise returns a TSLinearModel object (from tslinear) corresponding to cwe_name value (CWE ID)

• getNNAR(cwe_name) returns private attribute nnar when the argument cwe_name is missing; otherwise returns a NNARModel object (from nnar) corresponding to cwe_name value (CWE ID)

• getARFIMA(cwe_name) returns private attribute arfima when the argument cwe_name is missing; otherwise returns a ARFIMAModel object (from arfima) corresponding to cwe_name value (CWE ID)

• getBaggedETS(cwe_name) returns private attribute bagged_ets when the argument cwe_name is missing; otherwise returns a BaggedETSModel object (from bagged_ets) corresponding to cwe_name value (CWE ID)

• getTBATS(cwe_name) returns private attribute tbats when the argument cwe_name is missing; otherwise returns a TBATSModel object (from tbats) corresponding to cwe_name value (CWE ID)

• getStructTS(cwe_name) returns private attribute struct_ts when the argument cwe_name is missing; otherwise returns a StructTSModel object (from struct_ts) corresponding to cwe_name value (CWE ID)

• getPlots(model_list, row_no = 5, col_no = 2) returns plots of objects in the model_list with row_no number of rows and col_no number of columns arranged by marrangeGrob

• getAssessments(model_list) returns a data.table with columns cwe, method, MAE, RMSE, MAPE and MASE extracted from the list of models from model_list

• getAllAssessments() returns a data.table with columns cwe, method, MAE, RMSE, MAPE and MASE extracted from the models that the CVSSForecaster has built and stored in its private attributes.

• getBestModelList() returns private attribute best_model_list

• getBestAssessments() returns a data.table with columns cwe, method, MAE, RMSE, MAPE and MASE extracted from a subset of models returned by getAllAssessments(); subset represents the models which have the biggest number of lowest MAE, RMSE, MAPE and MASE values for a given CWE

• useBest(fcast_period, arima_stepwise = F, arima_approx = F, nnar_pi_sim = T) returns a list of lists, where each list has two elements: CWE ID and the forecasts fcast_period-step ahead into the future as a "forecast" object generated by the same type of model that generated previously the most accurate forecasts according to CVSSForecaster's method getBestAssessments

• plotUseBest(bestInUse, row_no, col_no, compact = F, actual) returns plots of objects in the bestInUse object previously generated by CVSSForecaster's method useBest with row_no number of rows and col_no number of columns; argument compact = T removes details from plots; when argument actual is not missing, then the actual values are added as a red line to the plot (should be the time series output of CWE object's method getTimeSeriesData)

• assessUseBest(bestInUse, actual) returns a data.table with columns cwe, method, MAE, RMSE, MAPE and MASE where the forecast accuracy numbers are calculated by using the forecasting output by CVSSForecaster's method useBest and the time series of actual values in a form given by CWE object's method getTimeSeriesData); example: CVSSForecaster's method useBest generated forecasts for 2018 based on the available data from 2013 to 2017, which was eventually checked when 2018 data became available

• setBenchmark(destroy = F) builds and assesses BenchmarkModel and assigns the result to the private attribute benchmark; if destroy = F then sets NA as the value of private attribute benchmark

• setARIMA(stepwise = T, approximation = T, destroy = F) builds and assesses ARIMAModel and assigns the result to the private attribute arima; the arguments stepwise and approximation are passed eventually to forecast::auto.arima; if destroy = F then sets NA as the value of private attribute arima

• setETS(destroy = F) builds and assesses ETSModel and assigns the result to the private attribute ets; if destroy = F then sets NA as the value of private attribute ets

• setTSLinear(destroy = F) builds and assesses TSLinearModel and assigns the result to the private attribute tslinear; if destroy = F then sets NA as the value of private attribute tslinear

• setNNAR(pi_simulation = F, destroy = F) builds and assesses NNARModel and assigns the result to the private attribute nnar; the value of the argument pi_simulation is passed eventually to nnetar as the value of that method's argument PI; if destroy = F then sets NA as the value of private attribute nnar

• setARFIMA(destroy = F) builds and assesses ARFIMAModel and assigns the result to the private attribute arfima; if destroy = F then sets NA as the value of private attribute arfima

• setBaggedETS(destroy = F) builds and assesses BaggedETSModel and assigns the result to the private attribute bagged_ets; if destroy = F then sets NA as the value of private attribute bagged_ets

• setTBATS(destroy = F) builds and assesses TBATSModel and assigns the result to the private attribute tbats; if destroy = F then sets NA as the value of private attribute tbats

• setStructTS(destroy = F) builds and assesses StructTSModel and assigns the result to the private attribute tbats; if destroy = F then sets NA as the value of private attribute struct_ts

• setCWEs(cwe_input) extracts start year, end year, end month, time series data and CWE IDs from argument cwe_input of class CWE; gives values to private attributes start_year, end_year, cwes, cwes_ts and end_month

• setBestModelList() finds the model objects corresponding to CVSSForecaster's method getBestAssessments output and assings the list of these objects to be the value of the private attribute best_model_list

realerikrani/nvdr documentation built on May 29, 2019, 8:06 a.m.