| 主权项 |
1. A method for advertisement placement, comprising:
obtaining one or more advertisement query keywords; determining, using one or more computer processors, in a multi-level advertisement information store, a selection of advertisement information for placement; and presenting the selection of advertisement information to be placed at a client; wherein:
the multi-level advertisement information store comprises advertisement information organized into a plurality of first-level categories, and each first-level category is associated with a respective plurality of subordinate levels of categories; anddetermining, in the multi-level advertisement information store, the selection of advertisement information for placement includes:
determining a first-level category based at least in part on the one or more advertisement query keywords;locating in each of the plurality of subordinate levels of categories associated with the first-level category, candidate advertisement information that corresponds to the one or more advertisement query keywords, comprising:
preassigning a correlation value between the advertisement query keyword and each piece of the advertising information, the correlation value relating to an extent of relatedness of a piece of the advertising information and the advertisement query keyword;determining, for the each of the plurality of subordinate levels of the categories associated with the first-level category, a predetermined number of candidate advertisement information based on the correlation value, the predetermined number of candidate advertisement information being two or more; andcomputing a comprehensive correlation value for the candidate advertisement information, the comprehensive correlation value relating to a sum of a first correlation value of a first piece of the candidate advertisement information weighted by a first weight and a second correlation value of a second piece of the candidate advertisement information weighted by a second weight, the first piece of the candidate advertisement information being different from the second piece of the candidate advertisement information, and the first weight being different from the second weight, the first correlation value being calculated based on a Pearson Product-Moment Correlation technique, the second correlation value being calculated based on a Pearson Product-Moment Correlation technique, the first correlation value of the first piece of the candidate advertisement information weighted by the first weight including the first correlation value of the first piece of the candidate advertisement information being multiplied by the first weight, the second correlation value of the second piece of the candidate advertisement information weighted by the second weight including the second correlation value of the second piece of the candidate advertisement information being multiplied by the second weight, and the first and second weights being normalized before being multiplied; andmaking the selection of advertisement information for placement based on the comprehensive correlation value of the candidate advertisement information, wherein the making of the selection of advertisement information for placement comprises selecting the candidate advertisement information from each category, the selected candidate advertisement information from the each category having a higher comprehensive correlation value than a comprehensive correlation value of another candidate advertisement information from the each category. |
