Skip to content

Latest commit

 

History

History
34 lines (27 loc) · 2.96 KB

weekl_report.md

File metadata and controls

34 lines (27 loc) · 2.96 KB

weekly report for pytz ipt 2018

The machine learning industrial train at cive at the University of Dodoma started on 30 th jully.This ipt deal with train student about machine learning and deep learning .But on the first week on the day one they was introduction of individual members of ipt and discuss which things shall curver on Ipt period.And also we arranged the time for study,presentation and discussion. On this Ipt shall plan study to understand machine laerning and deep learning and after thatevery member of this ipt to understand diffent algorithm and to develop project about machine learning or deep learning.

On the first week study about machine learning , the machine learning learnt bout how comuputer is programming with a data and perfoming a certain task.The machine learnig predict a data by applying the algorithm but not all the algorithm can solve the problem .The algorithm is used according to the natural of data .On machine learning can be classified according to amount of data or type data which will used on train.The data can be categorical or numerical.On this introduction of machine learning ,the machine learning divide into three namely supervised learning , unsupervised learnig and reinfoecement learning. But on supervised learning descover that the data which used to learning must be have inpute and output but on unsupervised does not has an output.

Apart from learning introduction of machine learnig their various chapter has studed about machine learning ,feature selection, model selection and model evaluation.On feture selection understand that the teature can to be select by eye instead of using either feature-base model selection of vizulazation of data.And also on model selection model must be selected according to the feature and the amount of the data.

on the model be select to understand more is Random Forest, the random forest is one of ensembles of decision trees on esembles of decision tree are the method which invole combination of model to form more powerful model.Random forest is best decause this model reduce the problem of overfitting by reduce anumber of n_estimato mean the number of tree selected.On this model does not well performing on high dimension,sparse.

On a middle of the week discussed about the project which shall do on the final of ipt.The project which discussed was General purpose Recommendation system eng.we discuss about the meaning of the project and tools which will use and time used to do the project.

Their some challenge face me during learning beacuse at the first time to learning machine learnig it take alot of time to understand the terminology of machine learnig.And also their several task on read for short time period.BUt after discover this challenge they decide to change the way of study from indivudual learnin chapter and divide charpter into groups of person and also in case of study we introduced presentation in older to understand more for ask question and provided answers.