mt_stat_evaluate function

Function to prepare and score data, finally predict current market type using pre-trained classification model

PURPOSE: Function that uses Deep Learning model and Time Series Column of the dataframe to find out specific market type of the financial asset it will also discard bad result outputting -1 if it is the case

mt_stat_evaluate(x, path_model, num_bars, timeframe)

Arguments

• x: * dataframe with one column containing asset indicator in the time descending order, typically 64 or more values
• path_model: String, path to the model
• num_bars: Integer, Number of bars used to perform transformation
• timeframe: Integer, timeframe in Minutes.

Returns

dataframe with predicted value of the market type

Examples

library(h2o)
library(magrittr)
library(dplyr)
library(readr)
library(lazytrade)
library(stats)

path_model <- normalizePath(tempdir(),winslash = "/")
path_data <- normalizePath(tempdir(),winslash = "/")

# start h2o engine (using all CPU's by default)
h2o.init(nthreads = 2)

data(price_dataset_big)
data <- head(price_dataset_big, 500) #reduce computational time

ai_class <- mt_stat_transf(indicator_dataset = data,
                      num_bars = 64,
                      timeframe = 60,
                      path_data = path_data,
                      mt_classes = c('BUN', 'BEN', 'RAN'))

# performing Deep Learning Classification using the custom function auto clustered data
mt_make_model(indicator_dataset = ai_class,
              num_bars = 64,
              timeframe = 60,
              path_model = path_model,
              path_data = path_data,
              activate_balance = TRUE,
              num_nn_options = 3,
              num_epoch = 10,
              is_cluster = TRUE)

# Use sample data
data(price_dataset)

# use one column for testing
x <- price_dataset[ ,2]

mt_stat_evaluate(x = x,
                 path_model = path_model,
                 num_bars = 64,
                 timeframe = 60)

h2o.shutdown(prompt = FALSE)

#set delay to insure h2o unit closes properly before the next test
Sys.sleep(5)

Author(s)

(C) 2021 Vladimir Zhbanko