<![CDATA[The search pattern is one of the ways that you can use to specify the search in the form of recommendations suggest that appears on the internet user's browser, by looking at how often the Web sites visited by the users simultaneously other Internet. Research on the hidden patterns that are found in the database into a field that is in great demand these days. In the hidden pattern, there is some useful information. Large-scale data will have difficulty when analyzed with simple data analysis so that it takes a special technique to be able to analyze large scale data, i.e. data mining. The results of the analysis will result in a sistem that would later figure out patterns of internet users accessing a website from the results of analysis of the association between websites that are frequently opened simultaneously, after knowing the pattern It then can be used as a recommendation sistem search. So, in this study using a sales transaction data to determine patterns of sales by using Association Rule algorithm and Modified-a priori from data mining. Association Rule is a method for finding interesting relationships hidden in data by using the calculation of the value of the support and confidence. Algortime Modified-a priori is the development of the a priori algorithms searching frequent itemset with joining processes (join) and pruning (prune) and produce a faster time efficiency using hash compared with the a priori algorithms. The technique used is the hash with the hash map. The data used in this research is the internet user search data retrieved from history browser internet users as much as 20 users with the number of transactions by as much as 300 transactions. The results of this research have the minimum value of the highest support i.e. 34.11% and generates the rule amounted to 2. Minimum confidence highest i.e. 80.00% and generates the rule amounted to 3. The length of the itemset that is formed are 2-and 3-itemset. Obtained as a rule that has a lift ratio of more than 1.]]>
