I am working on a problem to cluster various food items (like muffins, raw meat, sauces etc.) based on their names. In order to do this I am looking for a way to label different parts of food item names as either ingredient, food article or others. Eg: if the food item name is ‘chocolate muffin’ I want to identify ‘chocolate’ as an ingredient and ‘muffin’ as a food article. Similarly, for the item ‘Organic Carrots’ the labels should be ‘Organic’ as other and ‘Carrots’ as ingredient. I tried APIs for classification and parsing unstructured data, however most of them require words to occur in some context to label them (like in a sentence such as ‘I am eating a chocolate muffin’). My problem statement requires me to label them without any context simply based on the name for which a good pre-trained model is required. Any suggestions to tackle this problem would be appreciated.