hpipy.time_matrix.TimeMatrixMixin

class hpipy.time_matrix.TimeMatrixMixin

Bases: object

Time matrix mixin class.

The class provides a method to create a time matrix from a dataframe consisting of repeat transaction data. The method utilizes datetime data represented as periods and constructs a chronologically ordered matrix.

Methods

__init__
create_time_matrix	Create a time matrix from a dataframe of repeat transactions.

create_time_matrix(repeat_trans_df)

Create a time matrix from a dataframe of repeat transactions.

This function assumes the dataframe contains columns “period_1” and “period_2” that represent a pair of repeat transactions. The resulting dataframe consists of rows for each transaction pair and columns for each time period, with -1 indicating the first transaction in a pair, 1 indicating the second transaction in a pair, and 0 otherwise.

Parameters:

repeat_trans_df (pd.DataFrame) – Input DataFrame. Must contain “period_1” and “period_2” columns containing integer values representing periods (i.e., time series expressed as integer).

Returns:

DataFrame with columns ‘time_x’, where each row

represents a transaction pair and ‘x’ is a time period between the minimum and maximum periods in the input data.

Return type:

pd.DataFrame