The findings also confirm that the types and the distributions of dataset in continuous target are different from the categorical one; therefore using decision tree algorithms on the continuous target may be seen as a suitable candidate for crop physiology studies. These results are in general agreement with previous evidence. Within decision tree models, C&RT algorithm was the best for yield prediction in maize based on physiological and agronomical traits which can be employed in future breeding programs. One of the major advantages of the mentioned machine learning techniques for crop physiologists/plant breeders is the possibility to search throughput large datasets in order to discover patterns of physiological and agronomic factors. In particular, decision tree models are strong in pattern recognition and rule discovery by simultaneous looking a combination of factors in respect to yield, instead on analysing each feature separately. Another strength of decision tree models, which has a great potential use in agriculture, is its hierarchy structure. In a decision tree, the features which are in the top of tree such as ����Sowing date and country���� in decision tree generated by C&RT model or ����Duration of the grain filling period���� at decision tree with information gain ratio have more influences/impact in determining the general pattern in data, compared to the features in the branches of tree. Another example, in C&RT model, KNPE sits on the above of Mean/Max KW and has more contribution in dimension of target variable and possibly higher influence than Mean/Max KW. This topography/hierarchy structure of data in relation to target variable cannot be obtained from the current classical methods of analysis agricultural experiments whereas decision tree opens a new avenue in this field. As a pioneer study, this work opens a new avenue to encourage the other researchers to employ novel data mining approaches in their studies. Remarkably, the presented machine learning methods provide the opportunity of considering an unlimited wide range for each feature as well as an unlimited number of features. Increasing the number and the range of features in future data mining studies can lead to achieving more comprehensive view where this view is hard to be obtained from the separated small scale experiments. Recent progress in machine learning packages such as RapidMiner and SPSS Clementine, which offer a user friendly environment, Tracazolate hydrochloride provides this opportunity for the general agronomist/biologist to easily run and employ the selected data mining models without any difficulty. In conclusion, agriculture is a complex SIB 1893 activity which is under the influences of various environmental and genetic factors.