Ph-recognition

SREEVENK KOVVURI
3 min readOct 12, 2019

--

Ph levels are defined from 0–14 and each point has a colour associated to it. Colours are made up of RGB (red , green ,blue), so depending on scale of these colours we can get an estimate of ph-level or the colour corresponding to the particular ph-level.

Attributes of my dataset : red , blue ,green each having a range of (0–255).

The first thing which needs to be done is importing the libraries.

Read the ph-data.csv file in pandas to create data frame.

Its time to check the overview of dataset column information.

We need to check if any particular column value contains null values or not so that we can handle them either by using mean, median or mode values of the respective column or by dropping a particular column.

We don’t have any null values in our data set and now we just have to move on to next step.

Pictorial representation of my dataset

Normalization

As all my columns are in the range of 0–255 we must do min-max normalization.

Split the data set into 70–30 and train the model

In terms of accuracy : Random Forest>CART>Naive Bayes>Logistic Regression>Perceptron.

For Training Set -80 Testing Set -20

For Training Set -60 Testing Set -40

--

--