Abstract: Methods and systems are presented for classifying an item by dynamically probing different consumers for information related to the item. A description of the item may be parsed to identify a set of questions related to the item for probing consumers for additional information related to the item. The set of questions may be stored in a random forest associated with the item. The random forest may be traversed to select one or more subset of questions for probing a consumer for additional information. Once answers are obtained from a consumer based on the probing, the answers may be compiled with other answers stored in the random forest obtained from other consumers. The compiled answers may be used to generate an item classification for the item and may also be used to modify the random forest.

Abstract: Methods and systems are presented for classifying an item by dynamically probing different consumers for information related to the item. A description of the item may be parsed to identify a set of questions related to the item for probing consumers for additional information related to the item. The set of questions may be stored in a random forest associated with the item. The random forest may be traversed to select one or more subset of questions for probing a consumer for additional information. Once answers are obtained from a consumer based on the probing, the answers may be compiled with other answers stored in the random forest obtained from other consumers. The compiled answers may be used to generate an item classification for the item and may also be used to modify the random forest.

Abstract: Methods and systems are presented for classifying an item by dynamically probing different consumers for information related to the item. A description of the item may be parsed to identify a set of questions related to the item for probing consumers for additional information related to the item. The set of questions may be stored in a random forest associated with the item. The random forest may be traversed to select one or more subset of questions for probing a consumer for additional information. Once answers are obtained from a consumer based on the probing, the answers may be compiled with other answers stored in the random forest obtained from other consumers. The compiled answers may be used to generate an item classification for the item and may also be used to modify the random forest.

Abstract: Methods and systems are presented for classifying an item by dynamically probing different consumers for information related to the item. A description of the item may be parsed to identify a set of questions related to the item for probing consumers for additional information related to the item. The set of questions may be stored in a random forest associated with the item. The random forest may be traversed to select one or more subset of questions for probing a consumer for additional information. Once answers are obtained from a consumer based on the probing, the answers may be compiled with other answers stored in the random forest obtained from other consumers. The compiled answers may be used to generate an item classification for the item and may also be used to modify the random forest.