For this recipe, we need to analyze the trend of the data prior to building the model.

First, let's import the keras library:

library(keras)

In this recipe, we will use the shampoo sales data, which can be downloaded from this book's GitHub repository. This dataset contains the monthly sales of shampoo over a 3-year period and consists of 36 rows. The original dataset is credited to Makridakis, Wheelwright, and Hyndman (1998):

data = read.table("data/shampoo_sales.txt",sep = ',')
data <- data[-1,]
rownames(data) <- 1:nrow(data)
colnames(data) <- c("Year_Month","Sales")
head(data)

The following screenshot shows a few records from the data:

Let's analyze the trend in the Sales column of the data:

# Draw a line plot to show the trend of data
library(ggplot2)
q = ggplot(data = data, aes(x = Year_Month, y = Sales,group =1))+ geom_line()
q = q+theme(axis.text.x = element_text(angle = 90, hjust = 1))
q

The following plot gives us an understanding of the trend in the data:

Here, we can see that there is an increasing trend in the data.