**5. Research results**

HFO with scrubber, LNG and Methanol are the most excellent alternatives for the transformation of the industry. These three options are compared in the **Table 1**. According to todays and future expectations of price, infrastructure, regulation, availability, environmental impact, technology, capital and operational expenditures [13].

The technology used in the maritime industry has been bringing innovation to maintain safety and efficiency [14]. By this development, the data transfer between ship and shore became more available. More information onboard the vessels would help us to establish a higher degree of accuracy models.

Some machine learning tools were tested with real sample data from a ship which navigated from Norfolk to El Dekheila. The sample data used in the calculations are presented in **Table 5**.

In the pre-processing term for data cleaning, the columns "average speed, wind force, RPM, slip, swell" were selected to prevent overfittings of algorithms. Consequently, data science algorithms suit very well with these current sample data.

CRISP-DM "Cross-industry standard process for data mining" methodology which is given in **Figure 7** is one of the most common data science methodology [16]. When the procedure applied according to the CRISP-DM figure with the sample data, the model learned and predicted the columns successfully.

This model is based on correlations. Isabelle et al. (2013), draw our attention to the differences between correlation and causality, and difficulties of "Cause and Effect Experiment" [17].

By analysing data collected continuous data from onboard, it is possible to find causes of the events and prevent disasters as well as preventing climate changes, economic changes, epidemics, cancer. Ships are real-life laboratories for this methodology. Besides this, it proposes an evaluation methodology to take the right decision for company perspective.

Qualitative methods are mainly basing on expert's experience. However, when it comes to alternative fuel oils, the industry has not got enough experiences yet. We can say that; these qualitative methods are suitable for pre-AI shipping industry


## *A Review of Alternative Marine Fuels DOI: http://dx.doi.org/10.5772/intechopen.97871*


#### **Table 1.**

*Comparison of HFO - LNG- Methanol.*

**Figure 7.** *CRISP-DM Process [15].*
