update the readme

README.md

@@ -12,5 +12,3 @@ Experiments for the paper
 To reproduce the simulation results run `run_simulations.R`. It will print tables with FDR (false detection rate), power, and sample savings for experiments with both normal and non-normal data.
 
 To reproduce the results of the validation on the [Online Retail](https://archive.ics.uci.edu/dataset/352/online+retail) dataset, run `evaluate_on_online_retail.R`.
-
-Use `evaluate_on_real_world_data_2.R` to validate the method on real data. The script expects two or more input tables in the `parquet` format. The rows of the tables must correspond to events (in the terminology of the paper - for example, orders in the case of revenue monitoring). The tables should cover consecutive (adjacent) time periods of equal length and named in the chonological order. The `DATA_DIRECTORY` constant specified the location of the input parquet files. The `USER_ID_COL`, `METRIC_COL`, and `DTTM_COL` constants specify the columns of the tables with the identifier of the user that generated the event, the event value, and the timestamp of the event, respectively. The script saves the validation results in the `csv` format in `OUTPUT_DIRECTORY` and reports the false detection rates at the end of execution.