The development of an information technology currently carries a considerable impact against the pattern of life one on purchasing power. The current purchasing power are more likely to shop online because it's considered easier. But, how does a consumer know if the items to be purchased good or otherwise. Therefore it appears there is a review or comment on any goods sold. Review on items bring considerable influence against the purchasing power of consumers to know the quality of the goods, does not be surprised if a review into one of the main goals being viewed by consumers after the price. However, not all reviews provided the consumers can be understood by other consumers due to use the word is abbreviated, it use modern languages, in typing letters, the researcher proposes the creation of a system Analysis of Sentiment on the Reviews “Lazada†Berbahasa Indonesia Using the K-Nearest Neighbor (K-NN) and Repair Word Using Jaro Winkler Distance. Testing based on the value of precission, recall, and accuracy at each analysis sentiment without repair word, or with repair word. The test result with good accuracy value is present on the analysis sentiment with repair word is 76 %, with value of precission 0,76, and recall 1.
Copyrights © 2018