推移確率行列の作成

gistfile1.txt

# Adjust the code to exclude the last column (row sum) in the normalization process
# We only use the numerical columns except the last one for division
numerical_part_excluding_sum = filtered_matrix.iloc[:, 1:-1]  # Exclude the first (string) column and the last (row sum) column
row_sums_excluding_last_column = numerical_part_excluding_sum.sum(axis=1)

# Divide each cell by the row sum to normalize the rows, excluding the last column
normalized_matrix_excluding_sum = numerical_part_excluding_sum.div(row_sums_excluding_last_column, axis=0)

# Replace NaN values with 0
normalized_matrix_excluding_sum = normalized_matrix_excluding_sum.fillna(0)

# Re-add the station names as the first column
transition_probability_matrix = pd.concat([filtered_matrix.iloc[:, 0], normalized_matrix_excluding_sum], axis=1)

# 推移確率行列をCSVファイルに保存
transition_probability_matrix.to_csv(path + 'transition_probability_matrix.csv', index=True)

# Display the first few rows of the transition probability matrix excluding the row sum column
transition_probability_matrix