Chapter5_Splink_Settings.json

{
    "link_type": "link_only",
    "blocking_rules_to_generate_predictions": [
        "l.Year = r.Year and l.Month = r.Month"
    ],
    "comparisons": [
        {
            "output_column_name": "Firstname",
            "comparison_levels": [
                {
                    "sql_condition": "\"Firstname_l\" IS NULL OR \"Firstname_r\" IS NULL",
                    "label_for_charts": "Null",
                    "is_null_level": true
                },
                {
                    "sql_condition": "\"Firstname_l\" = \"Firstname_r\"",
                    "label_for_charts": "Exact match",
                    "m_probability": 0.9536693722036732,
                    "u_probability": 0.0037622388311744416
                },
                {
                    "sql_condition": "jaro_winkler_similarity(\"Firstname_l\", \"Firstname_r\") >= 0.9",
                    "label_for_charts": "Jaro_winkler_similarity >= 0.9",
                    "m_probability": 0.04185896073102217,
                    "u_probability": 0.0011643334683679213
                },
                {
                    "sql_condition": "ELSE",
                    "label_for_charts": "All other comparisons",
                    "m_probability": 0.004471667065304645,
                    "u_probability": 0.9950734277004576
                }
            ],
            "comparison_description": "Exact match vs. Firstname within jaro_winkler_similarity threshold 0.9 vs. anything else"
        },
        {
            "output_column_name": "Lastname",
            "comparison_levels": [
                {
                    "sql_condition": "\"Lastname_l\" IS NULL OR \"Lastname_r\" IS NULL",
                    "label_for_charts": "Null",
                    "is_null_level": true
                },
                {
                    "sql_condition": "\"Lastname_l\" = \"Lastname_r\"",
                    "label_for_charts": "Exact match",
                    "m_probability": 0.7039682117133251,
                    "u_probability": 0.0006134708934003129
                },
                {
                    "sql_condition": "jaro_winkler_similarity(\"Lastname_l\", \"Lastname_r\") >= 0.9",
                    "label_for_charts": "Jaro_winkler_similarity >= 0.9",
                    "m_probability": 0.2935505556049878,
                    "u_probability": 0.00029914541533406863
                },
                {
                    "sql_condition": "ELSE",
                    "label_for_charts": "All other comparisons",
                    "m_probability": 0.0024812326816870916,
                    "u_probability": 0.9990873836912656
                }
            ],
            "comparison_description": "Exact match vs. Lastname within jaro_winkler_similarity threshold 0.9 vs. anything else"
        },
        {
            "output_column_name": "Month",
            "comparison_levels": [
                {
                    "sql_condition": "\"Month_l\" IS NULL OR \"Month_r\" IS NULL",
                    "label_for_charts": "Null",
                    "is_null_level": true
                },
                {
                    "sql_condition": "\"Month_l\" = \"Month_r\"",
                    "label_for_charts": "Exact match",
                    "m_probability": 0.5013056458813202,
                    "u_probability": 0.0838861594985051
                },
                {
                    "sql_condition": "ELSE",
                    "label_for_charts": "All other comparisons",
                    "m_probability": 0.4986943541186798,
                    "u_probability": 0.9161138405014949
                }
            ],
            "comparison_description": "Exact match vs. anything else"
        },
        {
            "output_column_name": "Year",
            "comparison_levels": [
                {
                    "sql_condition": "\"Year_l\" IS NULL OR \"Year_r\" IS NULL",
                    "label_for_charts": "Null",
                    "is_null_level": true
                },
                {
                    "sql_condition": "\"Year_l\" = \"Year_r\"",
                    "label_for_charts": "Exact match",
                    "m_probability": 0.9571810981114466,
                    "u_probability": 0.019595044260833656,
                    "tf_adjustment_column": "Year",
                    "tf_adjustment_weight": 1.0
                },
                {
                    "sql_condition": "ELSE",
                    "label_for_charts": "All other comparisons",
                    "m_probability": 0.042818901888553355,
                    "u_probability": 0.9804049557391663
                }
            ],
            "comparison_description": "Exact match vs. anything else"
        }
    ],
    "additional_columns_to_retain": ["company_number"],
    "sql_dialect": "duckdb",
    "linker_uid": "heycobsb",
    "probability_two_random_records_match": 0.0001
}