Question

我是预测分析的新手。我试图从不同的地址提取国家代码（ISO）。我有一个包含所有可能国家/地区（dbname：Geo）的数据库及其各自的ISO代码。我正在尝试使用Rapid Miner或R作为起点（以及SQL）

例如：

地址：123 Main St 54321 US / CA.
结果：1）US 2）CA

制作预测模型以查找国家/地区名称的最佳方法是什么？

现在我只使用sql：

DECLARE db_cursor_country CURSOR
FOR
    SELECT  COUNTRY_NAME 
    FROM    TScotiaCountryUpdated
    --WHERE CHARINDEX(' ',COUNTRY_NAME,0) > 0

OPEN db_cursor_country;
FETCH NEXT FROM db_cursor_country INTO @Country



WHILE @@FETCH_STATUS = 0
    BEGIN   

    INSERT INTO TCountryNameMatch
    SELECT  FT.[Rank] ,
            TPartyNameAddress.PartyNameAddress,
            @Country AS Country_Name,
            CC_ISO
            ,CASE WHEN RIGHT(PartyNameAddress, 20) LIKE '%UNITED STATES%' OR PartyNameAddress LIKE '%NEW YORK%' THEN 'Y' ELSE NULL END  AS USFlag
            --INTO dbo.TCountryNameMatch 
    FROM    TPartyNameAddress
            INNER JOIN CONTAINSTABLE(TPartyNameAddress, PartyNameAddress, @Country) FT ON FT.[Key] = TPartyNameAddress.ID
            INNER   JOIN TScotiaCountryUpdated ON @Country= COUNTRY_NAME

            PRINT @Country
        FETCH NEXT FROM db_cursor_country INTO @Country


END;
CLOSE db_cursor_country;
DEALLOCATE db_cursor_country;`

RapidMiner - 来自地址的国家预测

0 个答案: