Predicting Percentage of Silica Impurity in Iron Ore Mining
Silica is regarded as the impurity in iron ore. With the help of some analysis and modeling of data, we can give a good approximation of silica concentrate which will reduce a lot of time and effort required for processing iron ore.
Data collection
Data was collected from a plant situated in the local region. Different columns of data representing information derived from the ore. Information parameters included timestamping, iron feed, silica feed, starch, and anima flow, and PH values of various chemicals.
Pre-processing
Normalization of the data was required for the dataset to process normally. Irrelevant features to the model were removed to enhance the lucidity in the application.
Implementation
Linear Regression, SGD (Stochastic Gradient Descent), and Random Forest were applied to the dataset in a specific fashion to obtain the highest accuracy.
Efficiency parameters
Efficiency parameters were applied to analyze the implemented model. The mean absolute error was 1% which is very less and the R2 score obtained was 99.84%.
Data collection
Data was collected from a plant situated in the local region. Different columns of data representing information derived from the ore. Information parameters included timestamping, iron feed, silica feed, starch, and anima flow, and PH values of various chemicals.
Pre-processing
Normalization of the data was required for the dataset to process normally. Irrelevant features to the model were removed to enhance the lucidity in the application.
Implementation
Linear Regression, SGD (Stochastic Gradient Descent), and Random Forest were applied to the dataset in a specific fashion to obtain the highest accuracy.
Efficiency parameters
Efficiency parameters were applied to analyze the implemented model. The mean absolute error was 1% which is very less and the R2 score obtained was 99.84%.
