E-COMMERCE PRICE PREDICTION MODEL
There are many machine learning models that are used to predict the occurrence of an event. One of the use cases in the machine learning model is predicting price of an item. There are several models implemented to build a machine learning model to predict price both for linear model and non-line...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/56135 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | There are many machine learning models that are used to predict the occurrence
of an event. One of the use cases in the machine learning model is predicting price
of an item. There are several models implemented to build a machine learning
model to predict price both for linear model and non-linear model. To choose
which type of model that is going to be used for price prediction will be based on
the dataset structure and also the outcome that is expected. For non-linear dataset, it is recommended to use non-linear machine learning model. In this experiment, comparison will be conducted between decision tree regression model and random
forest model to predict the price of wearable item in Shopee e-commerce. In this
process, not only does the model is being experimented, but also the improvement
of each model is also explored. In the end, the performance of the model is being
evaluated with the value of ratio between mean absolute error with dataset median
and also the value of ration between root mean square error with dataset median. In the implementation process of building the model, several steps are taken
which consist of dividing dataset with k-fold cross-validation and also doing
oversampling on the dataset. After implementing the model, the evaluation
process is conducted. Based on the experiment, it is known that with
hyperparameter tuning for random forest, it gives a much better outcome from the
value of ratio between mean absolute error with median which is 0.567 and value
of ratio between root mean square error with median which is 1.053 for overall
dataset. On the other hand, for the decision tree regression that also being
hyperparameter tuned, the value of ratio between mean absolute error with median
which is 0.901 and value of ratio between root mean square error with median
which is 1.426 |
|---|